{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Maschinelles Lernen (ML), Praktikum 2\n",
"Dieser Kurs vermittelt den Umgang mit den Pythonbibliotheken `matplotlib`, `NumPy`, `Pandas` und `Scikit-Learn`. Dabei werden Sie anhand eines Beispielprojekts vom Anfang bis zum Ende geführt. Im Rahmen dieser Übung werden Sie die folgenden Schritte durchlaufen: \n",
"- Einführung in Numpy, Pandas und Matplotlib\n",
"- Daten auswerten und visualisieren, um Erkenntnisse zu gewinnen\n",
"- Vorbereitung der Daten\n",
"- Modell Auswahl und Training \n",
"- Präsentieren Sie Ihre Lösung\n",
"\n",
"In dieser Übung experimentieren Sie mit realen Datensätzen. Hierfür stehen einige frei verfügbare Datensätze aus unterschiedlichen Fachgebieten zur Verfügung: \n",
"- [UC Irvine Machine Learning Repository](http://archive.ics.uci.edu/ml/)\n",
"- [Kaggle](https://www.kaggle.com/datasets)\n",
"- [Amazon AWS](http://aws.amazon.com/fr/datasets/)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"# Einführung in Numpy, Matplotlib und Pandas\n",
"\n",
"Numpy ist eine Python-Bibliothek, die für numerische Berechnungen verwendet wird. Numpy stellt hauptsächlich ein mehrdimensionales Array-Objekt zusammen mit effizient implementierten Funktionen zur Verfügung. Um Numpy zu nutzen müssen wir es zunächst importieren. Üblicherweise bindet man es unter dem Namen `np` ein."
]
},
{
"cell_type": "code",
"execution_count": 1,
"metadata": {},
"outputs": [],
"source": [
"import numpy as np"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Praktische mathematische Funktionen:\n",
"\n",
"| Operator | Beschreibung |\n",
"|--------------|----------------------------------------------------------|\n",
"| ``np.linalg.inv`` | Inverse der Matrix |\n",
"| ``np.linalg.eig`` | Eigenwerte der Matrix |\n",
"| ``np.matmul`` | Matrix-Multiplikation |\n",
"| ``np.zeros`` | Matrix mit Nullen erstellen (`.ones` für Einsen) |\n",
"| ``np.arange`` | Start, Stopp und Schrittweite |\n",
"| ``np.identity`` | Create an identity matrix |\n",
"| ``np.vstack`` | Vertically stack 2 arrays |\n",
"\n",
"\n",
"\n",
"Hilfreiche Funktionen für die Fehlersuche:\n",
"\n",
"| Operator | Beschreibung |\n",
"|--------------------------------|--------------------------------------------------|\n",
"| ``array.shape`` | Form des Numpy-Arrays abfragen |\n",
"| ``array.dtype`` | Datentyp des Arrays prüfen |\n",
"| ``type(stuff)`` | Typ einer Variablen abfragen |\n",
"| ``print(f\"Data type of integer is {name}\")`` | Einfacher Weg eine Nachricht zu erzeugen |\n",
"\n",
"\n"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Allgemeine Numpy Verwendung\n",
"Initialisierung mit Listen:"
]
},
{
"cell_type": "code",
"execution_count": 2,
"metadata": {},
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"(4,)\n(1, 4)\n(400,)\n(20, 20)\n"
]
}
],
"source": [
"array_1d = np.array([1, 2, 3, 4])\n",
"print(array_1d.shape)\n",
"array_1by4 = np.array([[1, 2, 3, 4]])\n",
"print(array_1by4.shape)\n",
"\n",
"large_array = np.array([i for i in range(400)])\n",
"print(large_array.shape)\n",
"\n",
"large_array = large_array.reshape((20, 20))\n",
"print(large_array.shape)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Listen mit verschiedenen Typen. Numpy verwendet einen Autocasts, der dem Array `from_list_2d` automatisch eine höherere Präzision zuweist."
]
},
{
"cell_type": "code",
"execution_count": 3,
"metadata": {},
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"Data type of integer is int64\nData type of float is float64\n"
]
}
],
"source": [
"from_list = np.array([1, 2, 3])\n",
"from_list_2d = np.array([[1, 2, 3.0], [4, 5, 6]])\n",
"from_list_bad_type = np.array([1, 2, 3, \"a\"])\n",
"\n",
"print(f'Data type of integer is {from_list.dtype}')\n",
"print(f'Data type of float is {from_list_2d.dtype}')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Numpy unterstützt viele Arten von algebraischen Operationen auf einem ganzen Array"
]
},
{
"cell_type": "code",
"execution_count": 4,
"metadata": {},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"array([0. , 0.69314718, 1.09861229, 1.38629436])"
]
},
"metadata": {},
"execution_count": 4
}
],
"source": [
"array_1d + 5\n",
"array_1d * 5\n",
"np.sqrt(array_1d)\n",
"np.power(array_1d, 2)\n",
"np.exp(array_1d)\n",
"np.log(array_1d)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Punktprodukt und Matrix-Multiplikation\n",
"Einige Möglichkeiten das Punktprodukt zu schreiben"
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"30"
]
},
"metadata": {},
"execution_count": 5
}
],
"source": [
"array_1d @ array_1d\n",
"array_1d.dot(array_1d)\n",
"np.dot(array_1d, array_1d)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Matrix-Multiplikation"
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"array([[170],\n",
" [390]])"
]
},
"metadata": {},
"execution_count": 6
}
],
"source": [
"weight_matrix = np.array([1, 2, 3, 4]).reshape(2, 2)\n",
"sample = np.array([[50, 60]]).T\n",
"np.matmul(weight_matrix, sample)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"2D Matrix-Multiplikation"
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"array([[19, 22],\n",
" [43, 50]])"
]
},
"metadata": {},
"execution_count": 7
}
],
"source": [
"mat1 = np.array([[1, 2], [3, 4]])\n",
"mat2 = np.array([[5, 6], [7, 8]])\n",
"np.matmul(mat1, mat2)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Elementweise Multiplikation"
]
},
{
"cell_type": "code",
"execution_count": 8,
"metadata": {},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"array([[ 0, 10, 20, 30, 40],\n",
" [50, 60, 70, 80, 90]])"
]
},
"metadata": {},
"execution_count": 8
}
],
"source": [
"a = np.array([i for i in range(10)]).reshape(2, 5)\n",
"a * a\n",
"np.multiply(a, a)\n",
"np.multiply(a, 10)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Plotting\n",
"Matplotlib kann zum Erstellen von Plots und Diagrammen verwendet werden. Üblicherweise bindet man es unter dem Namen `plt` ein. Die Bibliothek wird wie folgt verwendet:\n",
"1. Aufruf einer Plotting-Funktion mit einigen Daten mit `.plot()`.\n",
"2. Funktionen aufrufen, um die Eigenschaften des Plots einzustellen (z.B. Beschriftungen und Farben).\n",
"3. Den Plot sichtbar machen mit `.show()`."
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"data": {
"text/plain": "
",
"image/svg+xml": "\n\n\n\n",
"image/png": "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\n"
},
"metadata": {
"needs_background": "light"
}
}
],
"source": [
"# Import\n",
"import matplotlib\n",
"import matplotlib.pyplot as plt\n",
"\n",
"# Create data\n",
"t = np.arange(0.0, 2.0, 0.01)\n",
"s = 1 + np.sin(2 * np.pi * t)\n",
"\n",
"# Plotting\n",
"fig, ax = plt.subplots()\n",
"ax.plot(t, s)\n",
"ax.set(xlabel = \"time (s)\", ylabel = \"voltage (mV)\", title = \"About as simple as it gets, folks\")\n",
"ax.grid()\n",
"fig.savefig(\"test.png\") # Saves the current plot into a .png file located in the same folder\n",
"plt.show()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Plot mit gestrichelten Linien und einer Legende"
]
},
{
"cell_type": "code",
"execution_count": 10,
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"data": {
"text/plain": "
",
"image/svg+xml": "\n\n\n\n",
"image/png": "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\n"
},
"metadata": {
"needs_background": "light"
}
}
],
"source": [
"import numpy as np\n",
"import matplotlib.pyplot as plt\n",
"\n",
"x = np.linspace(0, 10, 500)\n",
"y = np.sin(x)\n",
"\n",
"fig, ax = plt.subplots()\n",
"\n",
"line1, = ax.plot(x, y, label = \"Using set_dashes()\")\n",
"# 2pt line, 2pt break, 10pt line, 2pt break\n",
"line1.set_dashes([2, 2, 10, 2])\n",
"\n",
"line2, = ax.plot(x, y - 0.2, dashes=[6, 2], label = \"Using the dashes parameter\")\n",
"\n",
"ax.legend()\n",
"plt.show()\n"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Pandas \n",
"Pandas bietet Datenstrukturen und Funktionen zur schnellen Manipulation und Analyse von Daten. Es setzt dabei auf die effizienten Datenmodelle von Numpy auf und analysiert Tabellen in verschiedenen Größen in Sekundenbruchteilen. Für das Maschinelle Lernen ist Pandes eine hilfreiche Bibliothek, da die Pandas-Dataframes und die Tabellenobjekte\n",
"des Frameworks grafisch als Tabellen aufbereitet werden. Üblicherweise bindet man Pandas unter dem Namen `pd` ein."
]
},
{
"cell_type": "code",
"execution_count": 11,
"metadata": {},
"outputs": [],
"source": [
"import pandas as pd"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Nun laden wir die Daten mit Pandas. Ums Einlesen kümmert sich ``pd.read_csv()``. Die Spaltennamen können mit einer Liste ``names`` per Hand festgelegt werden, falls die Datei keine zufriedenstellenden Spaltennamen liefert. Pandas bestimmt die Datentypen der Spalten automatisch, arbeitet aber schneller, wenn man die Typen als Dictionary im Parameter ``dtype`` definiert."
]
},
{
"cell_type": "code",
"execution_count": 12,
"metadata": {},
"outputs": [],
"source": [
"names = [\"sepal-length\", \"sepal-width\", \"petal-length\", \"petal-width\", \"class\"]\n",
"\n",
"# Optional\n",
"# dtype={\"sepal-length\": float,\n",
"# \"sepal-width\": float,\n",
"# \"petal-length\": float,\n",
"# \"petal-width\": float, \n",
"# \"class\": str})\n",
"\n",
"iris_data = pd.read_csv(\"iris.csv\", names = names)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Ein Blick auf den Datensatz\n",
"Wir verwenden den Iris-Datensatz, um die Klassifikation zu veranschaulichen. Es handelt sich hierbei um einen Datensatz, der die Länge und Breite der Kelchblätter (engl. sepal) und Kronblätter (engl. petal) von 150 Iris-Blüten aus drei Unterarten unterscheidet: *Iris-Setosa*, *Iris-Virginica* und *Iris-Versicolor*. \n",
"\n",
"
\n",
"\n",
"
"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Ein Blick auf die Daten\n",
"Schauen wir uns die ersten fünf Zeilen des DataFrames mit der Methode `.head()` an. "
]
},
{
"cell_type": "code",
"execution_count": 13,
"metadata": {},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
" sepal-length sepal-width petal-length petal-width class\n",
"0 5.1 3.5 1.4 0.2 Iris-setosa\n",
"1 4.9 3.0 1.4 0.2 Iris-setosa\n",
"2 4.7 3.2 1.3 0.2 Iris-setosa\n",
"3 4.6 3.1 1.5 0.2 Iris-setosa\n",
"4 5.0 3.6 1.4 0.2 Iris-setosa"
],
"text/html": "
\n\n
\n \n
\n
\n
sepal-length
\n
sepal-width
\n
petal-length
\n
petal-width
\n
class
\n
\n \n \n
\n
0
\n
5.1
\n
3.5
\n
1.4
\n
0.2
\n
Iris-setosa
\n
\n
\n
1
\n
4.9
\n
3.0
\n
1.4
\n
0.2
\n
Iris-setosa
\n
\n
\n
2
\n
4.7
\n
3.2
\n
1.3
\n
0.2
\n
Iris-setosa
\n
\n
\n
3
\n
4.6
\n
3.1
\n
1.5
\n
0.2
\n
Iris-setosa
\n
\n
\n
4
\n
5.0
\n
3.6
\n
1.4
\n
0.2
\n
Iris-setosa
\n
\n \n
\n
"
},
"metadata": {},
"execution_count": 13
}
],
"source": [
"iris_data.head()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Jede Zeile steht für eine Messung. Es gibt vier Merkmale: *sepal-length*, *sepal-width*, *petal-length* und *petal-width*. Sie Spalte *class* gibt das dazugehörige Label an. Die Methode `.info()` hilft, schnell eine Beschreibung der Daten zu erhalten. Die Funktion gibt eine Übersicht der Zeilenanzahl, den Typ jedes Attributs und die Anzahl der Werte, die ungleich null sind an. "
]
},
{
"cell_type": "code",
"execution_count": 14,
"metadata": {},
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"\nRangeIndex: 150 entries, 0 to 149\nData columns (total 5 columns):\n # Column Non-Null Count Dtype \n--- ------ -------------- ----- \n 0 sepal-length 150 non-null float64\n 1 sepal-width 150 non-null float64\n 2 petal-length 150 non-null float64\n 3 petal-width 150 non-null float64\n 4 class 150 non-null object \ndtypes: float64(4), object(1)\nmemory usage: 6.0+ KB\n"
]
}
],
"source": [
"iris_data.info()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Die Methode `.describe()` fasst die numerischen Merkmale zusammen."
]
},
{
"cell_type": "code",
"execution_count": 15,
"metadata": {},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
" sepal-length sepal-width petal-length petal-width\n",
"count 150.000000 150.000000 150.000000 150.000000\n",
"mean 5.843333 3.054000 3.758667 1.198667\n",
"std 0.828066 0.433594 1.764420 0.763161\n",
"min 4.300000 2.000000 1.000000 0.100000\n",
"25% 5.100000 2.800000 1.600000 0.300000\n",
"50% 5.800000 3.000000 4.350000 1.300000\n",
"75% 6.400000 3.300000 5.100000 1.800000\n",
"max 7.900000 4.400000 6.900000 2.500000"
],
"text/html": "
\n\n
\n \n
\n
\n
sepal-length
\n
sepal-width
\n
petal-length
\n
petal-width
\n
\n \n \n
\n
count
\n
150.000000
\n
150.000000
\n
150.000000
\n
150.000000
\n
\n
\n
mean
\n
5.843333
\n
3.054000
\n
3.758667
\n
1.198667
\n
\n
\n
std
\n
0.828066
\n
0.433594
\n
1.764420
\n
0.763161
\n
\n
\n
min
\n
4.300000
\n
2.000000
\n
1.000000
\n
0.100000
\n
\n
\n
25%
\n
5.100000
\n
2.800000
\n
1.600000
\n
0.300000
\n
\n
\n
50%
\n
5.800000
\n
3.000000
\n
4.350000
\n
1.300000
\n
\n
\n
75%
\n
6.400000
\n
3.300000
\n
5.100000
\n
1.800000
\n
\n
\n
max
\n
7.900000
\n
4.400000
\n
6.900000
\n
2.500000
\n
\n \n
\n
"
},
"metadata": {},
"execution_count": 15
}
],
"source": [
"iris_data.describe()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Eine Möglichket, nach nach sinnvollen Merkmalskombinationen zu suchen, ist die in Pandas eingebaute Funktion ``scatter_matrix``, die jedes numerische Merkmal gegen jedes andere aufträgt."
]
},
{
"cell_type": "code",
"execution_count": 16,
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"data": {
"text/plain": "
",
"image/svg+xml": "\n\n\n\n",
"image/png": "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\n"
},
"metadata": {
"needs_background": "light"
}
}
],
"source": [
"import seaborn as sns\n",
"\n",
"df = iris_data\n",
"sns.pairplot(df, hue=\"class\")"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Umgang mit Text und kategorischen Einträgen\n",
"Bisher haben wir uns nur mit numerischen Spalten befasst, aber jetzt wollen wir uns die Textspalten ansehen. In diesem Datensatz ist das die Spalte *class*."
]
},
{
"cell_type": "code",
"execution_count": 18,
"metadata": {},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"0 Iris-setosa\n",
"1 Iris-setosa\n",
"2 Iris-setosa\n",
"3 Iris-setosa\n",
"4 Iris-setosa\n",
"Name: class, dtype: object"
]
},
"metadata": {},
"execution_count": 18
}
],
"source": [
"iris_categories = iris_data['class']\n",
"iris_categories.head()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Es ist kein willkürlicher Text. Es gibt eine begrenzte Anzahl von möglichen Werten, von denen jeder eine Kategorie bzw. Klasse repräsentiert. Die meisten Algorithmen arbeiten lieber mit Zahlen, also transformieren wir diese Kategorie in Zahlen um. Hierfür können wir die Klasse `LabelEncoder` von Scikit-Learn verwenden.\n",
"\n",
"**Notiz.** [*Scikit-Learn*](http://scikit-learn.org/) enthält effiziente Implementierungen vieler Machine-Learning-Algorithmen. Es bietet einen guten Ausgangspunkt verschiedene Algorithmen auszutesten, ohne jeden einzelnen zu entwickeln. Einen Kurzüberblick über Scikit-Learn finden Sie in dem Ilias Ordner."
]
},
{
"cell_type": "code",
"execution_count": 19,
"metadata": {},
"outputs": [],
"source": [
"from sklearn.preprocessing import LabelEncoder\n",
"lb_make = LabelEncoder()\n",
"iris_data[\"class_code\"] = lb_make.fit_transform(iris_data[\"class\"])"
]
},
{
"cell_type": "code",
"execution_count": 20,
"metadata": {},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
" sepal-length sepal-width petal-length petal-width class_code\n",
"count 150.000000 150.000000 150.000000 150.000000 150.000000\n",
"mean 5.843333 3.054000 3.758667 1.198667 1.000000\n",
"std 0.828066 0.433594 1.764420 0.763161 0.819232\n",
"min 4.300000 2.000000 1.000000 0.100000 0.000000\n",
"25% 5.100000 2.800000 1.600000 0.300000 0.000000\n",
"50% 5.800000 3.000000 4.350000 1.300000 1.000000\n",
"75% 6.400000 3.300000 5.100000 1.800000 2.000000\n",
"max 7.900000 4.400000 6.900000 2.500000 2.000000"
],
"text/html": "
\n\n
\n \n
\n
\n
sepal-length
\n
sepal-width
\n
petal-length
\n
petal-width
\n
class_code
\n
\n \n \n
\n
count
\n
150.000000
\n
150.000000
\n
150.000000
\n
150.000000
\n
150.000000
\n
\n
\n
mean
\n
5.843333
\n
3.054000
\n
3.758667
\n
1.198667
\n
1.000000
\n
\n
\n
std
\n
0.828066
\n
0.433594
\n
1.764420
\n
0.763161
\n
0.819232
\n
\n
\n
min
\n
4.300000
\n
2.000000
\n
1.000000
\n
0.100000
\n
0.000000
\n
\n
\n
25%
\n
5.100000
\n
2.800000
\n
1.600000
\n
0.300000
\n
0.000000
\n
\n
\n
50%
\n
5.800000
\n
3.000000
\n
4.350000
\n
1.300000
\n
1.000000
\n
\n
\n
75%
\n
6.400000
\n
3.300000
\n
5.100000
\n
1.800000
\n
2.000000
\n
\n
\n
max
\n
7.900000
\n
4.400000
\n
6.900000
\n
2.500000
\n
2.000000
\n
\n \n
\n
"
},
"metadata": {},
"execution_count": 20
}
],
"source": [
"iris_categories = iris_data[\"class_code\"]\n",
"iris_data.describe()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Klassifikation\n",
"Für den Anfang werden wir die Klassifizierung vereinfachen und lediglich versuchen, zwei Unterarten der Iris-Blüte zu erkennen. Dieser Ansatz ist ein Beispiel für einen *binären Klassifikator*, mit dem sich genau zwei Kategorien unterscheiden lassen (z.B. *Iris-versicolor* und *Iris-virginica*). Erstellen wir also zunächst einen Datensatz mit den beiden Klassen *Iris-versicolor* und *Iris-virginica*."
]
},
{
"cell_type": "code",
"execution_count": 21,
"metadata": {},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
" sepal-length sepal-width petal-length petal-width class_code\n",
"count 100.000000 100.000000 100.000000 100.000000 100.000000\n",
"mean 5.471000 3.094000 2.862000 0.785000 0.500000\n",
"std 0.641698 0.476057 1.448565 0.566288 0.502519\n",
"min 4.300000 2.000000 1.000000 0.100000 0.000000\n",
"25% 5.000000 2.800000 1.500000 0.200000 0.000000\n",
"50% 5.400000 3.050000 2.450000 0.800000 0.500000\n",
"75% 5.900000 3.400000 4.325000 1.300000 1.000000\n",
"max 7.000000 4.400000 5.100000 1.800000 1.000000"
],
"text/html": "
\n\n
\n \n
\n
\n
sepal-length
\n
sepal-width
\n
petal-length
\n
petal-width
\n
class_code
\n
\n \n \n
\n
count
\n
100.000000
\n
100.000000
\n
100.000000
\n
100.000000
\n
100.000000
\n
\n
\n
mean
\n
5.471000
\n
3.094000
\n
2.862000
\n
0.785000
\n
0.500000
\n
\n
\n
std
\n
0.641698
\n
0.476057
\n
1.448565
\n
0.566288
\n
0.502519
\n
\n
\n
min
\n
4.300000
\n
2.000000
\n
1.000000
\n
0.100000
\n
0.000000
\n
\n
\n
25%
\n
5.000000
\n
2.800000
\n
1.500000
\n
0.200000
\n
0.000000
\n
\n
\n
50%
\n
5.400000
\n
3.050000
\n
2.450000
\n
0.800000
\n
0.500000
\n
\n
\n
75%
\n
5.900000
\n
3.400000
\n
4.325000
\n
1.300000
\n
1.000000
\n
\n
\n
max
\n
7.000000
\n
4.400000
\n
5.100000
\n
1.800000
\n
1.000000
\n
\n \n
\n
"
},
"metadata": {},
"execution_count": 21
}
],
"source": [
"iris_data = iris_data[iris_data.class_code<=1]\n",
"iris_data.describe()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"In einem weiteren Schritt erstellen wir den Zielvektor $y$ für diese Klassifikationsaufgabe. Bei einer überwachten Lernaufgabe sind das die Labels in dem Datensatz."
]
},
{
"cell_type": "code",
"execution_count": 22,
"metadata": {},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"0 0\n",
"1 0\n",
"2 0\n",
"3 0\n",
"4 0\n",
"Name: class_code, dtype: int64"
]
},
"metadata": {},
"execution_count": 22
}
],
"source": [
"# Sie können eine Variable mit Punkt-Notation extrahieren\n",
"y = iris_data.class_code\n",
"y.head()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### Auswahl von Merkmale (engl. Features)\n",
"Die Spalten, die dem Modell als Input dienen (und später zur Erstellung von Vorhersagen verwendet werden), werden als Merkmale (engl. Features) bezeichnet. Vorerst werden wir ein Modell mit nur zwei Merkmalen erstellen. Die Auswahl mehrerer Merkmale erfolgt durch eine Liste der Spaltennamen. Jedes Element in dieser Liste sollte eine Zeichenfolge sein (String). \n",
"\n",
"## **Aufgabe 2.1:** \n",
"Wählen Sie zwei geeignete Merkmale. Verwenden Sie dafür die oben beschriebene Scatter-Matrix. "
]
},
{
"cell_type": "code",
"execution_count": 23,
"metadata": {},
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"['sepal-length', 'sepal-width', 'petal-length', 'petal-width', 'class', 'class_code']\n"
]
},
{
"output_type": "display_data",
"data": {
"text/plain": "
",
"image/svg+xml": "\n\n\n\n",
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAYEAAAESCAYAAAAbq2nJAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAgAElEQVR4nOy9Z3hc2Xnn+Tv33spVAKqQIwEwZ3YzdWKzk1uh3S0rR0uWZMmS1+Px+rHX9ti73t15PLafcdr1eLyWZ2zLClYObkktqdUK3a2OJJs5E0SOhcrxprMfbhEECJAEgSIJAvX7QAKoOqdu3fTe84b/K6SUVKhQoUKFlYlyuzegQoUKFSrcPipGoEKFChVWMBUjUKFChQormIoRqFChQoUVTMUIVKhQocIKpmIEKlSoUGEFUzECFSpUqLCCqRiBChUqVFjBaOWYRAhxH9A5fT4p5b+WY+4KFSpUqHDzWLQREEJ8DlgNHAas0p8lUDECFSpUqLDEEYuVjRBCnAI2yYr+RIUKFSrccZQjJnAcaCrDPBUqVKhQ4RazYHeQEOJpHLdPCDgphHgNKF56XUr51OI3r0KFChUq3EwWExP4i4UOFEK8Gfj90q/rgU9LKb+1iG2pUKFChQoLoBwxgT+XUv7e9f52jfGvAo9KKTOL2pAKFSpUqHDDlMMIHJJS3n3F345KKbfNY2w38DfXcx3V1dXJzs7ORW3ncqG3t5fbsS9000YRgsms4/HTFEFt0HPLt2M6t2tf3EpyukW6YAAQ9Gh4XCq2lLjVmeG8O31f6KZNPKcD4HWpVPtcGJaNEAJNETc0152+LxZD0bSJZXWklAQ9GudOHo1KKeuvNWYxMYFPA78OdAshjk57KQT8fJ7TvAP45lXm/yTwSYCOjg4OHDiw0E1dVuzateuW74ufnhnnjf4EHk0h5NWIZnR2dNTw8PqGW7odV3I79sWtZixV4GsHB5FSsn9dPT8+PYEtJQ9vaGBHe83U++70fZEtmnzx1X6yuskvbGpEVQTPHBtFCHjn3W20R/zznutO3xeLoW8yyx996zi6afPe3e28c2d73/XGLCYm8EXgGeBPuezfB0hLKWPznONJHEMwCynlZ4DPAOzatWtBy5XO3//uVV/r/bMnFjLlsmQ8XeCFs1HqQx72ra1DiJlPXsm88yRaNG3etbkJn1sl5HXdjk1dcTRWefn4A130TGT4yZkJRpJ5Gqu8JEpPzcsFn0uls87PZEanudrH2bE0AFJCqrQSqnB9LFtS5XVRMCzUea6gFmwEpJRJICmE+F+ufE0I4ZJSXvPICSGaAF1KObnQbahQHl6+MEl/LEd/LMfaxiDN1T7Aufmbls3+dfWMJQtsaA7RUOW9zVu7vCkYFvGcTlOVd8oYFwyLH50aw7IlhmWzsSlEZ22AgmHhdam3eYvLw2A8z6H+BIZlU90zySMbGxhPFfBoKhuaqma9XzdtJrNFGkLeed/sVgJSSnTTJKdbKOImG4FpHALagTgggBpgRAgxDnxCSnnwKuPeBny7DJ9fYZG01PjomcgS8Di+WHDcEF9+fQBbSrJFk6ODSV65OMmWlhoiQfdt3uLliWHZfPHVfpJ5g62t1Ty2qZGBWI5vHBri7FiGuqCbXZ0RPG6Vb74xRMir8eF7O3Frd74EmKoKTgwlKZo2axuCDMRy9ESzKEKwrb166sHkEl89OMB4qkh3fYC37Wi9TVu99BhJFni9N4Fl27x0ITqvMeUwAt8Hviml/AGAEOJx4M3AV4D/Duyda5CU8h/K8NkVysDuzgjddQECHo2L0SznxjP4XAoXJjJYtiRTNAHIFi3G04WKEbhJFE2bs2NpknkDr0tlVa2f7x0fJZYtsro+wLa2Gh5aX8/XDw0CkC6Y5HXrjjUCZ0bTnBlLs6OtBpDU+F3kihYuVTCeLiIlWFJOuYguYdmSaNpxh42nileZfWUyGMti2TaWlIwmC/MaUw4jsEtK+alLv0gpfyiE+C9Syt8WQtze9JEK80ZTFEzb5ocnxrClJKeb2FJi25JHNtRzfjxLV12ADc2zl+YVyoeUEltKpJR899gImYJJtmhy7+o6Hlhbh6Yq7F/XwKsXJ2kL+6j235mxGcuWfP/4KLaUjKcKvGmzEwzWNAUQ3NUeZixVwK0qrG8KzRirKoJf2NTI2bE026cFxytAfbUXt6qg2zZN83TdlsMIxIQQvwd8qfT7e4G4EEIF7DLMX+Emc7AvzvNnJwh6VAIelXTBpC3iJ+hxYUvJnq5aPnxv1+3ezGWPR1NY2xgiXTBZ0xDk+8dHGUrkWd8U4s1bLiuzNFV773gXiCIgEnARzejUBT1Ued2saQihmzYdET9jqQL9k3lUBeI5nYbQzBvappYqNrVUHkiupDbgIehzYVo2DVXzewYvhxH4APDHwLdwYgIvlv6mAu8pw/wVbjKD8RwAmaLFzs4aBmN59q+rx+/WMGx7xgV4fjxNz0SWHR01sy7MCgsnUzR5tWeSLS1V1Fd5uRh1aic3NYeoX4b7WQjB7q4Ih/ri7FxVQ7XfxS/fu4ps0aS52sfPzkxwMZpBUQQjiULlXJsnIa+L5ioP6YLFqkhgXmMWbQSklFHgP1zl5fOLnb/CjSGlJG9Y+N0zD23BsHCpypyZFPd011I0bMJBF2/0JbAl/OjUOB+5r3PG+4qmxXePOkv4sXSR9+1ux5YSj7Y8MlRuJy+ei3JqJIWUko5aPwOxPC5NocbvZt+6OnTTxq0pmJaNacs7PivIsiU/PD5KwbR59uQ4H3ugiyqvi6pS6rGUkmzBRFGFo1AG5HULj6agVLKBrspwIs94uohpSc6MpeY1phz9BNYBv8PspjKPLHbuCjfOd46OcH48w8bmqikXwonhJM+eHKPK6+L9ezrwuWfeQBqrvLxndzu2LemfzJPKG4S8s08NTVGm3EUCyT++0INtS37prlbawvMv5qkwmyqvhi0lp0fSDMbzCAHtET8Pb2jg+bNRnj8b5fFNjfzkzDgFw+aJbc2srg/e7s1eMIqAi9Ecw8k8m+aIM0kgo1soCggFDvTGeOGcU8vyvt3taOqdGQy/2RimxUiygGU7abfzoRzuoK8C/x/wP7jcVKbCLaRgWBzqi1Pl0+iZyALQE70sxdQzkUVKJ+8/miletfpSUQT71tRxqD/OfatrZ72uKoL37elgNFkgkdN54ZyTgtYfy1WMwCLZ0V5DTzTLcCJPS42PsVSB7W3VGJaNbjqhtSODSbJF5xLrm8ze0UbAsiVBr0qwoBHwzF7VqArU+F1oikBKyYUJ53yeSBdJFUwigUqG2lyMJItoqkAgKRrzux2XwwiYUsq/L8M8S4KrVRkv5Qrjn5+PcnQwCcDG5hCjyQI7OsJTr+9cFSaR04kEPLTU+OacwzRtEPDDk6Pops2zJ8f40D2rkJIZy++gR2NNQ5C8btE7mcO0bDa3VN/cL7gMkVLO2Lcvno8ykS5i2pLRVB6vpnF4IMkjG+pprvaAhEfW1/P8uShZ3WRb252ZFWPbEkURKEKgWxLTsjGsy4IApmmjaQqWdArCTEUggT1dtTx/doK2sI/wHZoRdStorvLgmACbGv/8DGU5jMDTQohfx9EAmt5P4JrSEUKIDwMfwQkgf1BKOVSGbVmRXPIPK0KwqzNC3RXCbi01Pn753s6rjn/m+Aife7mPuqAb25ZMZHQ2NIX4ny9epGjavP2u1lnGw+dWedfOtrJ/l5VApmjyldcHyBsWT25roaPWj6d0DLvrgmxrq+bwQILz4xn6Y1kuTGSJBNzs6AjzS3fduVlBRwYS/OTMOC01Pt5xVyuGZVMwnRhHIqfzh988RiJn8PF93dQF3WxpdR4uqn1uuuoCdNXNL9C5kilaNnnDwrYlk9n5SYuUwwh8pPT/7077mwS6rzZACNEK7JdSPlqGz1/x7FoVZiyVp6naN8sAXIvToynyusWL56JYtmQkWSDidxP0aOimRbrgFIldmMhcdQUBTsDu+HCSpirvDQl9rVSG4nl6ohl00+b0aIqOWj8PrKmjIeQh7HcT9rs4Ohgnp5vkDYhndXwulVcvxmbUaVyMZpnMFNnatnRWYufHMyTzOltba2YVsTmBb+f7x7I6DUEPHk0h7HdxcjhFNOPctF66EOU/vWUj/bEcXk2hqy5ALKPz9NFhNjaH2NM121VZweHMaAbbltgSxlK3KCYgpVxIAvmbAFUI8RxwEvgtKWUlnrBAXu6ZpG8yT99kntX1QRrnUSTSG83yzLFRADpr/YwmCzRWecjrNhOZAn6PRmOVl6JpzandMp0fnRrj/HgGVRF89P7OirjcdVCFU95vmDa65fj7VUWwsXSD/+efX+SlCzGSeYM1DUHaI37CfhcPrb+sCBzL6nz78BBSOj8vBcZSBZ4+MgxAqmDOUpm9qyNM6uw4bWE/dUEP966p4/RIil2dYdY3VtFZF2AyU+TxTY28MZDg3JgTB1hVG+Bzr/RxZjTND06M8tfvDczrHF+JuFSwpfMUblrz090sR3aQH/htoENK+UkhxFpgvZTyO9cY1gi4pZSPCiH+HEdH6BtXzDtDSrrC1bkkNCaEU6gxvzGXf97eHuYTD67GsiWfeb6HDsO56Vi2jWU71auX+P7xUS5MZLh3dS13l+IO0+e6UoG0wmw8LpUtpThKjc/F1w4OMpEu8vjmRlbXB6eEv6p9Lj6wp4P71tTxem+M7xwdYUNTiEc3NiKg5PuVS2afC5xzQcq5z8P1TaEZ1b/719Wzf91lw/bn77zcguTHp8c51B9HVQSPbGiY2ieC2XPrps233hgiltN5y5aV3e486HWhKs4x8Hvm9zBWDnfQPwMHgftKvw/iZAxdywgkgZ+Vfv4xsOvKN5RDSnqlcP/qWqp9Lmp8ruuqfCZyOn2TOVY3BGmp8RLPGtzd7tyQVEXwrp1t9MeyWLbk5+cdgdczY2kaqrzkdYtTI07u8ZGBxJQReGxjIy01PhqrvAQ9N35KFQyLs2Npmqt91IcWpjSSLZqcH8/QEfETXsKZI9miSSyrs64xRMFwJCFOj6QIeDSODyVZXR/kg3s6nJx5n8Z9a+oAONyfQDdtjg4meXBdPeGAm3fc3Uo0U5wzMD8Qy5HMG2xsrrplKpsNVV72dkUYSRa4t3u2yyZdMPjhiVE2Nlex6TrJBG5V0Bb2oQiBEPBbj63le8dG2TiHku1IMs9QwnF9HB9KYdqSo4MJ1jaEZqVDL3faIwFCHo2CabO5NTgvhc5yGIHVUsr3CiHeDyClzIvrP5q8BHyi9PMO4GIZtmPFoqnKjAYjV0NKyVcODJAtWjx9ZJhjQ05GUdCj8qFS4Lg+5KE+5CFVMDg+lEK3bNY2OE9vXpfC2sYgF8azbG29fBF7XeqUQVgI3z8+ysVoFrem8Kv7uhZUfPb0kWFGkgX8bpVP7OtesgVF/35kmLNjac6NpQl6NPxujUzRZGNzaCpfXtMU3n73zADwltZqXrsYY31TCFcpR7494p8zBjOeKvD1Q4NICYmcwQNr627+F8NJ33z1Ygwp4bXeGPvWzmxo9dfPnuXEcAqXKvib9+2gLnj1B5Z1jSGOD6dwqQqragPOqmjv3B6BpmovDVUeEjmDjc0h4lmd506Nc3Yss+KSFzIFg4IpsWzJaOLWBYZ1IYSPUl2fEGI107KE5kJKeVgIkRdC/BSIAn9dhu2ocB2kdGoGhhJ5AtOekM6MpfnU5w7QUuNj35o6LkSz7GivIejVKJYqVcFx9fzitpZFfL7kuVPj9MVyPLi2jrWNjnExSn5x23bSJq+GbUt+cGKUkWSBhzfM9DdfmsO0JRInD/2Z4yNMpIs8uqGRc+NpeidzPLCmbpYg2c3guVNjc36eadlTgbtUwWAwnierW1i2zX2rnZt1NFPkmWMjeF0qT25vwetSuXd1LffOUbsxF+a0/WhYTrvB7x4bwaMqPLm9haFEjp+djdIR8fPYxoayuZOKhrNSLBg27ZHZiQQT6SIjyQI+l4ppSf7wG0d5vTfOm7c08duPr5/x3oYqL5/av3pen+vRVD64d9XU75miydNHhtnbFVlxRiCZL1Io1QdEM/NTWC2HEfhjHDnpdiHEF4D7gV+53iAp5e+U4bMrzAPblvREMwQ9LgJujfqgh/aIj7DfTSynE00XiecMJrM6iZyO360xGM+R0y0My6az1j/rqW4hpPLm1Orj9d74lBF405Ymjg8maY/4rymHEM0WOT3qdJw62Bef8dovbmvh1EiKrvoAqiIYTuSnAos/Px9lNOXI6r7WG7vpRiBVMKbqNr5/fISQV5vKrnpsUyNuTeGe7lpe7Znk+FCSkWSekYRwblzdtRwfSnJuPIOmCDa3VN+wUFpLjY+3bG0imTPY0VHDaxdjRNPODeHCRIZjQ0lSeYPjQ0l2d4bnnU9+PUxbUh/0kDcuy1v3RrO4NIXWGh9b26pJ5B3BON2yePbUGAXD5ttHhmYZgUtjNVXccCGiZUvcmsLEPG+Cy4lzY44OmARit8oISCmfFUIcAu7Bidn8x5Ke0LLiTm5V+fMLUQ70OkG2dU1BhhMF2iM+R54AQWOVl55olvqgh4JhM5TIUO3TODWSRgJbWmrKYgSCXo2WGi/DiQLrGi9Xu1Z5XVO+72sR9rupD3mIZoqsbZhZLRsOuGfMURt0Uxt0E8vqbG6tQlUFQ/E86xpufpVt0K3RWuPjYH8cy3JccO/e1U5rjY8XzkUZThRIe01qQ25MW1IwbeJ5g2BJqqNg2JwbS6OpCsoC1RGmZ3R11QU4MpDApSq0hX0UTYvRZIHmam9ZM7kaq7x01wdJF0w2NFVxfMiRKwGnT/DOVRGyRYumai+NQQ+ZgkXBtJhLAWL62Hfc3cqq2vnXCPhKjer3dkXK8r3uJHwucUlqiaI1PxHnxTSav/uKP42U/u8QQnRIKQ8tdO4KC+eN/jgnhlNsaAzyzIkxUnmDjoif40NJfG6V9Y0h8oZFyONCSieYtqszwscf6Matwj88f5GCaTGZKdJc7fhsb8Rb8NL5KBcns9zTXTtL1kBVBO/Z1Y5hyes2QkkXDH5wYgxNEbx5SxNel4pLVfjg3o6rjjcspx9CfyyHqjh+5fft7sCtKWxtrZ7X55YDRRG8e1cbTVUeDvYnGE7k+fwrvQTcGufHM9QFPbxwbgJFCFbXB9EUJ0NrW1sNL56L8vrFSVY3BIn43WgLiG3ops1f/vA0E2mdTz7YzYbmKj61fzVCCFRFsHNVhK2tNbhUUdbMIrem0FTlxaPp1PhdnBhKcnI4haoI4jmd+qCHSNBFc7UXGwWvS8G0LPyu2behnG7N+fN88LgU6oIeuu9gWY2FopuX/anzPbaLWQn85TVek0BFQO4WI6Xk+bNRbCk52B9jOO64QOI5naBXw6spvHpxkpDXRdGweHBdHTndYldneCoY++T2FnqiGTY1h/jmG8OkCybv2dV+zc81TZsDfXHqgm5evegUir90YXLKCNi2ZCCeozboIejRcGvOyXly2HGZzJUpcnwoxUDMWdqeHk1PBb6FEFPjr6Q3muXIYIITw0kiATcjyQKmZfPIhgaiGZ3z4xl2dUZuiSEQQnDP6joQgu8czTOZ1jmaTLKxqYpYtohlS4QCprT52L7V+NwqG5pCfOm1AXxulVTKZHt7zYL0gV69OMkrPTEMy+abbwzyB82bZgiu6brFd4+PsL29pqw3ysF4jjOlBvEHe+P4PSohrzYVpH+5J0osYxDLJFjfGCQScCMEc2aE3dVRw1iqgEdTWN94Y+67eM7gUH+MgmHy6MbGxX+xOwg57dK4ymUyi8U0mn94oWPLzbVcNSsJIQSrav1cjGa5q62GeHaComFzT3eETNEi6HHEukaTRVbVBtm5avZyubMuQGddgOFEHt2UuFSF/ljumn7pv/3JOV7pieF1KexbW0+6YLJqWtbKj06NcWI4hd+t8iv3d+LRVH56Zpy//+kFhIDfeHgND1zhbmoL+9AUgaIIWqrnVxikCMGpkRTxrI5bFZwcTnF8KMnJkRRHBpJkiiY72qv5g7dumuceXRxuTeHBdfVEM0VODKeoNS2qfS7etLmRv/vpBYqGzVPbW6duVJYtqQ95ODGcxLQlx4dSbG6ppmme3/8SXk3QH8uV5BiMWa//1lePcGQgjtel8pVfu5faG6gyvxZ1JSOf1U06av0EPRqddU6MpjXsI6dbDCcK1Abd1IW87OyM0BvNzplZdmEiw/nxDEI4DWRuJC5QMCxGkkUyhWsq1yxLdneG+fwr/UiclfBL8xhTjsDwFEKIz0gpP1nOOSvcGE9tbyGjm4Q8Gm+/u428btFQ5SVTNPFoCqoQU69fi2TewC6lmCRy1041u9TntWA4T911Ic+ULjw48gYnR1JUe10UTRuPpjKccFYpUjKV4z2d9oifX93XjRDMXztfwLa2amwbqn0qw4lhJjM6FyYyZHVHAmM8feura39pRyuPbWrEpTjuF69L5W+bqyia1ow0SVURvH9PBz86NcrJ4TS2lCTzxg0bAcNyYiKGKQnMcZzH086+LxgW0UyxbEYg4NH4yH2d6JY9VS/yiWnHsC7oYUtrFT6XiqYq/MnbtjCWKdA8R21LPOsYr0tprm03mIGswFQfgpXE5pYafnnvKjK6wZu3NPMv8xhTViPAHEVfFW4tiiKmbsAhr2sq8De9iOvS66mCgWHac94E1jWGiGaK6KbN3auufQV+7P5OPvtyL5taqmmp8ZHMG4Q82pRPUhECtypwaQJZilU9tb2FiUwBgeDJ7ZfTTguGRTJv0BDy3HChT3ddgPvX1JEpmIT9Lp4/O0k8p7OjLUxtyM2xwSRvvw0CbNOPiWU7DcAjAfecQVlVEexf14AqFHxudVYA/GpICaPJAvUhD03VHuqCHlJ5Y85MqN9903r+8fketrfVsP46kiA3iltTZrjbrjyG07+zpim01lx+ws8WTXK6RX3Iw10dNeQNE01RpuQ05otHU/C4FO5ds/I0hlbV+tm3vp5Ytsg9cxTszUW5jcB4meercJOYSBf50mv9mLbkLVubZukDqYqYd0bQaLpIbdBLumDyuZd7yRQttrRW8wubHDfH2sYgmaJJtc+Fv6QdH/Rq/MdH182Yp2hafP6VPtIFk7tXhWdICswHIQT3ra7jp2fG+e6xEc6MpfFoCr2TWT76QBdPbb/9CpzfO+Y0/akPefjg3o45g3del8pjm27Mlx3L6fzba/2sbQzSXRtgNFnAsOw5G4vs7apl7xITYUsVDD7/Sh9Fw2b/+nru7gjzyIaF+fNtW6IpgvH0yksRTeYNTo2k0E17qgfD9SirEZBSvrmc81UoD8mcwUsXokQCbvaWng5iWR3TdtbLY6kiG25QcsW0bF44H8W0JBejGc6NpfG5VeJZnVhOR7esKSOwsTnEcCLPusYQf/fjcxweTPKrD3TNigPkipeVS8dKef0L4chAgpfOR0nlddyawrGhJP/2Wj8hr8buzgjpgsGpkTTb2qpvKPVwOgOxHIcHEqxrDN1Q3cGl7xXNOMFhTS1Pdo5ZSgccSxWp9qpkCga6ZTNWagD0med78LpUPrV/Necn0nztwCCbmqt41652zoymOTvmBN/LqQJbMCxeOBfFpToPFNeSr0jmDIqG8x3GF3HsAQxbki5a9JSC1CuJyUyRC+MZiqbNmnmuIheTIvo01/C6SSmfusbYTuBV4BSgSykfX+h2LAVudSOaRE6nyuuaUxrBtiWpgjHj9Z9fiHKmVGTVFvHTWuNjTYOjW583LHZexd2T1y0sKWe4kkzLJlM06Z/M8XpJIiBnmGSKJhLJQDxLwZCcH8+QK/nhf3YmyljKCY7++PQYihD81bNnZxmBcMDNA2vrGIznpqpn54ss+c/dqsLJkRSxrI5hA6akbzLLVw8MsL2thmTOmDKAI8k8n3xwflWpV/LsyTGSeYOeiSyr6wNoqoJuOlru1T7X1H66shDr0Y2NvNEfZ11jqKwtEqt8LlbV+rm7I8wbfTFyuollw2iqwFcPDPBaKWuruy7Az85O0FMqGtvVFeb7x52+0ePpIh9/oIuiaVE07RlxnfmSLhikCgatNX4O9cc5XioOdOIBV9cLagv72LkqTDynz9uNcd1t0eeXJ7+8EBQMi5xhTcX0rsdiVgJ/sYixAM9KKT+0yDlWHImcwT//vJfu+gBv2zHbvfGtw0P0TeZY2xickni41IrPrSlTvYNVRVwzfS6aKfLl1wcwLclTO1roqgtg25KvHBhkLFUg6FE5PJDAlhKvS+HCRAavJkgVLAqGxViywP98wZGEupQC2BDy4ndrFAyLpuq5+xPs7oywu/PGi3x+eHKMQ31x+mM5Tg6nSBdNBI4PPpEzyOlp1jeGqA16kDjusEhg4QHR2qCbZN6gxu9CVQR53eILrzqurH1r6zg1miaaLrJzVZgHp7m1blZzFJ9L5R13OxIJPz8/Tt5w5DMG4jncmsJIsoAiIORVMSyb0VSBoEcj5NGo8buIZXXqgm7SBYMvvtpP3rB4bGPjNW/cVzKWKvD7Xz9KTrd4/56OKV++IsR120EKIWbsp3KwAuPC5AyTk6MpDNNmMHaT+wlIKX92/Xddk4eFEC8A35BSztIOWg5S0jejyviS/vylHPrpSCmnfMD9kzkO9MbIFE3u6a6lPeIn5NXm/XQ3ksjTM5HBsiUvnB3nr34YZ3V9cOrCGk8X2dJShQ2cHUtTH/RgS/C4bCSOz/+Su6mzFLCNBNy84+4WjgwmePxG/U/XYSCWm1Lo9GgKfrcG0sSWToFY0KuxrjHI45saMWyb8VTxhrNupvPE1mZGSoFYIZxiqAvjThZSJOAiVspuGYjPPk43g3hO59e/cIj37mrjYjSPUpJ0tmxJd32Q/evq0FSF2qCX3Z0RNEUQ9rtRFYX37m5nIu0UBw7G81PFWYPx3A0ZgQvjGcZSBSxbcmQgzi/d1UrY70JTlUpP4FvEZEbHoyooQDJ/iwTkSv0D/hTYBExdVVLKq3YWw6kuXocjNPdtIcRzUsqj099QkZKem5BXo6naO6VHPx0hBPvX1XNyJEVtwD3VCB7goSsafFwPibO0N23Jl14fIJEzONgX59P7V6NpjmrpufEMlm2zf209Xz7Qz7rGEORjHPUAACAASURBVAG3xvHhFE9sbSarm0gJW1urp1IVAx6N9kj5n4QfXFeP16XSVO3jYP8k6YJJypZ4NAUQ3NVewwf3rEJRBG6hUB/yTKlxLgRNVWb4z1UFsrpFVrfQFIW93RH6JnNzSirfDMZTRQ73x7kYzfBf37mdrx4YoGhaPLK+ge3tNYwk8nhcKmsbg1T7nNVLe8Q/5a669F3aI342t1SRyBvsusEVWUvYWenldHMq1nI9afObyXwzi5cTrTU+gh4XecNkVd384jvl6ifwxzhKoA8DH+U6vU2klEVKSqNCiO8AW4Cj1xqz3FhoHMHnUnn/nquvjLa317C9vYbxVIFTI06u+SUX0I2QLZqcHXNWAh5NIV0wcKkK96yOsLbRWeZfWu5/8dV+JtI6isjyZ+/c6jyF32LWNYZYV6os/YOv5zhBCkUIXKrApalICRndIuJW+c5RJ0NnegbTYvG5nZWGYUnaIn7uW13HfQsLNywIKSXpglFagVloqoLEMebJnMFYqohLU8gVLdojft53lXNIVQSPb17YKq3K6+bBdfVYtix76ulCUMTNrwxfatT4nWNgS8m6xvkdg3JcrT4p5XNCCCGl7AP+z5Kb54+vNkAIEZJSXgrd3w/87Xw+qJyVwcu9yrihysv797ST0y06F+CDHksVCfvd2FLSXO2hPeKnIeThwkSWZ0+N85bNTXSVJAfeGIiXxhQYjudZc4Nl/uXk2GCSgXiOxiovYb8bj0uhUDQZSuQ5N55mTyAylTp3fjxTNiNQ5XXxgb2riOd0uhaYcbSoz/e52LkqzI6Oas6NZzFtR7K6bzJL72QW05aYusVQIn/Tmu5EAm4+sLeDVN5YEk3hl2hLiZtKfcjD+/e2ky1adNbeupVAQQihAOeEEL8BDAHX8z3sE0L8Z5zVwItSylfLsB0VrmAxS/GH1tfzcs8kRcPm3TtbOTeeocbv5l9f6sW0JWdHM/zN+3YATuHXF17tZ3V9kO76yxf/pbaUN7P9oZSX2yv2TWb50akxagMeCobF/nUNBFyCbxweQVMFLTW+qVqCE8NJ7lpEI5y5iATct833XeVzsX99A7s7I8SyBZBg2jY1fjdbWqrpj+XwaOq80wYXSl3QKVS7nSilNqt1C+xSd6fTEPLCDTyHlcMI/BbgB34T+M84wnEfudYAKeX3gO+V4bMr3CRq/O6pnq8nh1Mk8wkEEMsZpAsmgWkuphqfm41NVY46pHRK9mNZna8dHMCWjozwQttGXotU3uD/ee4cd3U4hWUuVUEIyOpO5enLPZO0hb3s6YqUMpMcJ/Gergh7lpnMcNCj8dH7uwB44WwRw5ZIBLrlyDU76bNO6ue8ZTjuWIQTGF+JGaILYNFOMynl61LKDJACflNK+Q4p5SuL37QKC2EyU+T13hjxbPk0ck6NpLClZCxdZGtrNVtaq9g9TXzu0utDiTyJvJMVczGaJVu0yOsWPfOsXLxR8oaFlI6RAqeZyt0dYRJ5J34xmswT8rpoqfbx7l1tZdPIWeqMpor43Sp+t0rRsDk7lqZvMsfFaHbOrLJlh5QoQlAwb0yCeqVSjuygXTjB4VDp9yTwMSnlwcXOvRJZbOHZ1w8Nki1aHB9KTj0ZLpbt7TVEM0WaqgNsb6/hYjQ7I+tlR0cN8ZxOS42PSCnbZE1DkBPDSaRkKmBbbgJurdTf2JGZHojlONgXpyHkYTCWZ01DkLYaL49vbprq7LUSeGRDA18+MEAia/D2u1sxLMlIIn/LGs7fblRVQVUE3UsgLnEnUA530D8Bvy6lfAFACPEAjlHYVoa5K9wgopSYtdDL/blTY5wZS7OttZqLkzl00+bJ7c38Wqnf6w9PONWl03N2p2fmXKLa5+LDpeb1N4ugV+PTD11OwUnldX5wYpRUwWBzczVP7Wjhia3NS7bp/E1FTv1D2O+ain/cqCjfnYginMYy5eyatpwphxFwAf8khGjAufcIYOU8di0x3rmzjZ6JzIICgLppT/XG/cmZcUxLYtmS0yMpGkJe8rrFiZLr5Y3++FSjl9vN+fEMUkqGEgXcqgISEvki58czJPPGTcuGWUpYtuToYII1DUF+fHqckaRTNPjtw8P8w4d24VZV3JqyoCY1dxqWLRFA/2T2dm/KHUE5jEAz8APg73EePd4LxC+1n6y0mby1OBkqCwt6ujWF9U0hzo1lWF0X4vsnRrBsSdF0Imxel8KahiA9E9kbqiS9mZwdS/Pdo05n0z1dYVrDPlRF0FkbpCPip8q3Mp4GY1md506Nc3I4xUPr6/nsS72kCwaPb2pCUQRb25bG8boVKEJgS2heQS7AxVAOIyBwUkKvrAu4j0qbyTuOt25tRm5xBODG0gWklFT7nCdpIRzt/+lpmbcb3bycAhLwuPj7D+3Etm2EKG//3KXOJfecYdk0Vfv4zm/uw7IsVHX5u3+uJBxw85F7O2gJV2IC80HIeSrNzRooxDtKP+4HmoBvUaoCBpBSfmPRW1eirq5OdnZ2lmu6BZHIGRRNC1UR1AY9C/a5L5be3l5u975YKlxrX1hSMpnRkVLid2sLqpq+k7iR8yKVN8gbFooQ1AU9LDdbefp8D+GGFjyaSo1/ZawEr8bBgwellPKaWaCLuTKeLP3vBe7Cqfx9FqgG6oGyGYHOzk4OHDhQrukWxOdf6WMiXURVBL+2v3uqMfutZteuXbd9XywVrrUvxtMFvvBKP+AErp/Y1nwrN+2WcyPnxdcODjIQyyEEfPyBrmUXQO1Yv5X/9b99nYYqDx/cu+p2b85tRQhxXXf8YlREP1r6kJeBPwL+UEr5USGEBpxZ6LxLlTdtbuLoYILOusBtMwC3mlvdJ6GcNIS8/MKmRibSRXZ1lrcy+E7n0Q0NHOqP0xb2LzsDAE5m2ra2ara1LY3EhaVOOdbI26SUXxFC/AGAlNIUQtzRj10T6SLPHB/B51J5cnsLXpdKfchzTf39CkuH8+NpfnY2SlvYx+ObGldUbACcuMB3j46QyOm8aUsTzVf0bggH3Mv6XNZNm97JHE3V3ptSqb7cWExnsXtxgr+qEOKPgHohxG/jSErfEQXbyZyBx6XMKqM/PpRkMuNU3PZOZmf1313p3Iw+CeXkQG+cVN7gWLbIhsYQq1ZA0ZDEyRAK+10MxvNcjDrpkUcGErOMwHInp5uMpwq8djHG5jkk1yvMZDGyEW4gCKRxmr/U4egIPQW8b/GbdnM5MpDgn35+kc++1EumaM54rbs+gKYIgh5tRVWaLhfWNoYwLZuL0SxfOzjIwb7Y7d6km048q/PZl3r5wYkxGqs8Uz0DbrZg3FLEsGwODyTIXXFdV5ibxXYW+5kQYhz4AtAOvBPYCwzPZ47SyuEdUsoHFrodC2Uo4RTT5HSLeFaf0Ud3VW2ATz+0GkWIWdWmUkrOjKVRhLhpcggVbpxM0eTcWJqOiJ+dq8LUh9x87cAgQggG43l2LvP4oFHqODcYz+F3azyyoYF4VqerbuUZAVVRaA/7cLtWXj+BhVCORvP7gbcAEWAjcAF4DrhmNE4I4QG2L/TzF8uergjZokkk4KZ1jqf9qzUBPzGc4tmTY84v226eLk6FG+PpI8OMJgv43Cqf3NdNe9jPXavCTKSLt6y71+0k5HXRUuNl56ow46kC3zo8VGqkY7JvbXl79y51bCmJ5w0Ms9KQcD6Uo9H8PUAep2Xky0ArMJ+mwL8KfBb4v+d68Wb3GK4Lenj3rvYZf/vpmXGeOzXOvrV1V+2udKlvLoBpVU6ypYJpS6SUnB/L8N1jIzy8oYGHp7XUPDyQoG8yy+7OyLJ08fndKu/d7VwnI8k8l8p/pp+vlzg+lOQrBwbY2Fx1zS51dyqWLZlIFUgXjNu9KXcEi240L4RQgQTwGPAJHIMwea2xQggXsF9K+XdCiDmNwO3oMfwvL/VOSR8/trEBRZm9GthWkktQBGxsrqwClgpPbmvmRyfHyeoW58czBD0aD29wjECmaPKT0+MAZIsWH9i7/G5802mu9vHEtmaSeYPtc6RJfvalXvpjOc6MpnlgbR3t4fl1oLpTMCwbw5L0RCvaQfOhHCmiE8Bh4C+klAkhxB7gepG4Xwa+WIbPvmE+/fmDHB9K8uC6Ov7k7TOFTpurvPREszRWeWcZgAO9MV66MMnq+iBv3dq04tIOlzo1fjcPb6hnNJXHsCRhv4svv97PmdE0li05MpjErQke29jIc6fGODGcYueqMPevqbvdm152CobFaxdjJPMG9UEPmaLJX/7wDB5N4X9/chOqAkPxPNV+F9We5VcnUDRszo5nMCr9BOZFuTqLfQZ4b+nG2ImzIrgW64EdQohPAZuFEP9BSjmvPsOLIa9bHBlIYNmSV3ou26lkziBnmPynt27gpQuT7OmcLcB2bCiJZUvOjqV5ZEPDipDkvdOoDXr48H2dFAwL05J8/eAgJ0aShP0eXKqgI+LHoyocG3L6HBwdTC5LIzCSLHBhIk1etzg9mqInmiWnW+R0i9cvxtnSUk22aFIf9GKz/FyaEokqJMmKO2heLNoISCm/L4RYC2wo/em0lLJ4nTG/d+lnIcSLt8IAgKOlHgm4GYjnWFfjuHLiWZ0vvNqHYUks20ZVFCazxiyXwfb2Gl6+MMnq+kDFACxhqrwuqrwuzo6meObEKHndoj1ssbMzQmOVl7tWhZnM6pwcTi0ZKexyY1gWL5yNUjRtWsM+Ht3QyKG+OB5NYW9XhO+fGGUya6AqCh5t+WXQSAmG5WQJVbg+i8kOekRK+eNpQnKXWC2EmLeA3GLTQ5N5g+NDSToiftojc/s2pZQcG0qS0y0e29iIhClBsWTewCgFeEeSBdrCfmLZ4iylzLs7wtxd5sbkFRaHYUlePBdlQ3NoRnPzTNHkRyfHUYWgxu+iodrLn7x964yx04PGy41YxgABQjjCh5taqvkfH9k99XrArbG7tNrVLbnsmn+oiiDk1ahahpIYN4PFrAT2Az/mspDcdCRlFJC7Ft8/PsJwosChvjifeLB7ziba58czPHfKCQx2RPy4NWXqKXBVrZ+9XRFSBYPHNzXSE82ysbmq4vO/A0jkdF7vjXF2LM3HHrjcSvPZk6MULZuWGh8hn8ZvP7buNm7lrcelOn5xw7KZSyT4oQ31vHYxRnvYT/Uy7LegCIEQEAksv+92M1hMdtAfl/7/aPk258ZxlfL5VVWgXOXGrZs2Pz0zjmlLdrTXkCmaZHUTcQHCfjePbWycKgrb3FpNIqfzlQMDBNwaj29unPqMCkuLS4fbVXJpFHSTv3j2LC+dj5LTLfxujce6Irx6McZ4ujjjOC9nFKGQKZoYlk3BMBlK5Ph/nzuP16XyO4+voyHk5Re3tcw51rIlz54cI5nXeWRD4x2pvWPaklTeZLhUEFrh2iz67iaEuCCE+IIQ4lNCiE3l2Kgb4a1bm9m3to737GzDXboZ2LZNIqdPvef8RAZbSmxb8rMz4wzF83z+5T76J3McG0pOVQ8DxDI6B/tiDMXznB1L0zMxvzSzgmFhz5GTXeHmEfa7eXRjA2+/qxWAF89HOTaYZCxVIJHTGU/mePFclNcuTnJkIEFPNAM47sG8vrQzR4qmhWktTILr/EQGpI2QkvGUzrcPD9MzkeHUcJIfl1JlEzkd05w9/0Asx6mRFMOJwpKT25jvNXapB/Zo+pqhyWWL0w1w/ud3ObKDNuFIRewD/kIIsQE4IqV8exnmvi6H+uK8ejFGz4SPd+1sw7Ylv/eNowzG8zy+qZGP7+sm5NEYjOexbEm1T2Mkmacu6OFAX5wqrwtfyYX058+c5lB/nNYaL6tqg7g0QWPV9Z+EDvbFeP5slPqQh/ftbr9qtfFS5VqCcEsZVREz5ILXN1bhdanYEnJFE1PCwb44miI4Ppwip5s8vrmZiUyRC+MZtrRW8wublp6aZt9kln8/PIxbU3jf7g6qb7QximURzzs3gf5YhrdsbWY0WUBTFVprfPzry7189+gIzdU+/uwdW/C6L98G6kIeAh6VnG5dNcZ2OzjQG+OFczd2jekrsGI4UzT50mv9ZIsWT2ybu+D1SsphBCzAKP1vA2PAeBnmnRfnJ5ynu6FEnrxhkcobDMadJ/vDgwkAUgWTthofpmUT8rlorvZhS0lb2IeqKORKT4UnRpwm6yPJIv/XU5txaeqcMYYruTDurBYm0kVSBUeK4hK2LbkwkaHG774jl9Z3Eu21fv7bB+7id75ymGNDScZTRRRFoAgQUjIYy3N6NEky7wiLnR/PLEkjcDGaxbQlpm4xlMiTKhhICR2187spv9afnPo5ljVoqPLwps1NqAK8LpU3+p3rYiSZZyxdJBKQDCXydNYGCHo0fuW+LnTLnqGndbs5O5ZmMlOkYFizrrEKlxlNFhhNFiiaNhfm6cUox1FOAceAvwL+UUp5zWrhcrO3q5aXL0Tpqg8S8GgEPBr3ranl9Eiat213/J5rG4LYOBbK61JJ5A266wK01Pio8btpqfEC8MTWZn5yepz71tQR8s3/JNvVGSZ31qQ17Cd8xVPbi+ejHOyLoyqCD9+7ihp/5eS9mYS8LtY2hjg3niHkc1Hl1Qj7XWR1i4Lp1A/ct7qOE8NJ7lqi2V7b2moYSuSnVqhfOzgIOK7P9U3Xr1L/8D3tPH1kGMOyeXhjA1taqhmI5XFrCmsagrxtRwtffn2AtQ1BWqu9/MvLfaQLJq1hH+/Z1Y5bU6Zcq0sFw5IMxPNEAm6889y2Kt/SMWK3Cr9bYSiRp2BYzLd1cDn20vuBB4BfB35VCPES8LyU8rkyzD2DgmHxcs8kXk1lb1cERRFEAu7SzdzFf//peRQEv7ave8YSVxGC9U0hLEuiqdAWDlDjd/HL93bOmP+9uzum9Ffmw6mRFAOxHDtXhfmV+7vmfM+lVYZl27zSE8OtCe7trruhWoOTwykG4zl2zVHEVsFhNFngcH+MEyNpTgynWNvopI3+xiNrONQX54VzURqrvFT5XOzpirCna+nuS79bpaXGh8+lUjBM+mM58rqJW1NwqYLu+msrg4aDTjOVbNFkTX2IcMA9o+7lofUNPFRKkbVsScFwztHbLb18sC9OIqdzT3ctLlXh5Z5JXIrgnu5aQl6NHe01CAHmHDe3S2P3dteiKQKPJuic58ppOWFaTqp7TjdJFvTrD6A8xWLfBr5digW8BaeC+H+D8qcfH+yLc7i0lK0NulnXGOLZk2OMpQp85cAAyZzhGIagm/dMF4cTElUIpAL3ra4jHHCxZZHNJjJFkx+cGEVKJxf7Pbvb53zfvrV1eF0KRdPm5HDK2RzElK7N9UgVDH540vmcZH7hFZB3qt9/vjxzfIQjAwle742hKgKfW6Uh5OHlC5P0TeZwqQrtEadwaqlzoPfyed5dG0ARzoUd9Gb5ztERfv2h1df0iX/l9X5Gko5L9Euv9/Nr+1df9b2qInhqeyvnJ9Jsab19DVgG4zmePzsBONk91T4Xh/riAFT7XTy60Sl4aw37ZuX/DyXyM8ZatnS0g+bpDllOOEktOWwJz5+JzmvMoo2AEOLrwA7gPPAi8GHg1cXOOxeXDr4QTPkrk3mdQ31xJBKEk/kxFM/xj8/3kC4YPHN8lPqgh52ralAUhfVNobKc7C5V4HM5AbRrLTsDHo2H1jcwkS5yeiSNLeWs9xd0k//yzGlGkwU+dn8X96yuJZ7V+c7RYaYnQ1Qtw5zuclHldeF3q2iqglsVNIW8fP/4KN88NEh9yMvG5ioeWtdAeAn7kqV00jNf7pmkaNg0VHmo8mvEswa6aYOUnB/P8LmXe3liWwsNVd6psYPxHI/95c9489YmOiN+bOlUzvrmEdPqqPXPO95ws3CrCmfH0mSLJqvrgyQx+N6xEVRF8PD6eqp9rqs+OAXcKpoiMG1JldeFBGzJgrOr7mQCHpVk3sCWYMyR/TUX5XAH/RlwSEp5U3PuopkiG5tDKIpT8XhJDtjrUumqCxDwaOxbV4sAnj87SV8my/Nnx9EtSbZo8q5dbWxormJT89VbRdq2JJbTCfvdqKV88mROJ5rV6a4LMJnVqfa5cKkKHk3lXTvbuDiZ5a72y77lTNEsZSHNvGHXhzx88J4OckVrxgWXzBkcH05yajiFJSXfOz7CPatrOTWSIlpqcbm7K0JztZfO2gB/WK4duoxIFwweWl/Pjo4antrWSqpg8OPTY7zeO4ktBZMZneZq75RrbikxmSliS0mqYFIf8nB0MImmCBSX4NGNDUykinTW+eio9dEQ8pIpmiTyJi9diPLWrS1TvvtM0SJV0PnB8RH+jyc301bjJVO0uKfLOTeTeQNNEQRuc7B3MlMk6NXwaCqmZZPIG0T8bvKGRX3AhUcTqAr0RrMIAQK4MJFhxxXxm+lja/xuPnjPKlJ5g1XTri25AlO2J9I6bkViWMypgjwX5XAHvX7pZyHEZ6SUn1zsnFeSzBt87uU+NEVgSYlLVXjv7nbqgh42NVeRyBl0RPzc211Hvmjyp987w2S2iKYIsrpJwK1yZjRNfyyHW1Wu2nLv6aPD9ExkaQv7ePeudmIZnd/92hEyRZPV9QHqQ17qQh4+uKcDS0qePjJMPGeQLVrsX1fvuKVeH8CW8OT25lm+27qgx2nIWeLUSIofnBhFNy2SeYOcblG/3skg6qoP8MZAAk0RbGquqmRDXIXhRJ6vHRxESnhkQwMHB+J879gwI8kCugWqIvG5VfrjefyepaX59MK5CZ4/O8FPz0xQ43fxxNZmBuM5TgwnKZo2z5+bZHt7NYYlWd8U4i2bm3ju9DinRlIk8zrxnMGH7lmFS1WQUpLIGdT43XhUhWjGwLRtYjmTCxMZnj4yjKYI3rO7nYaQ9/obd5O+74HeONU+Fx+6ZxXfODTISLLAusYQm1uqeOlijLxusbYhRFa36JvMIZj7if5rBy+PfWJbM5GAe9Y1svRM/s1nNJknU/IavzHPOo9yPxbsKvN8gFPxC84Tgdel4lIFE+kidUEPe7tr2bkqPOUjTesWPrdKq8uHbkneuqWZRMFwCkikE0C8mhEYThQAx/8qpWQwkWMyU8SwbE6NpDFtZ1WhWzZF0yaec/Z232SWQ/0aiZzOWLqAZUtOj6aI5wzWNQYJldxYPRMZMkWTzS3VqIpgMJ5jNFnAsGz2dIWp9rmnmqI3V/v41P7VCFgRVa4LZTTl7G+Ac+NpdNMmmtExLUnE7+J9u9uRODICt7pAzLYlJ0dSeF0KaxpmZ/UMJ/KkCwapvA5S8p1jw6yKBIgE3PRN5sjrFkPxPN31QR5cW09nXYCPP9DFv702UCqIM8gVLar9zsp0W2sV4YCH/niOoEfFlgqZosVIolASVZOMp4q3zQhcquBN5g3SBYOfnB7j1IjT06A24Cbg1nCrCrGcQV43CJWM9lhqZoDTsiWjqcKMOediJQrIHSjFUQAS+VsUGL6Cm1IfEPK6aA37CAdcvHxhEk1RcE27MU4PkjVWeXliWzPHhpI8vL6ebNGiocpDXrcomjY7Oq6uHPnIhgaODCbYVNIOaqvxEvBoJPIGNT4XsayOS1VQFUG1z8V9q2vpj+XIFk1+dmaCWFZnPFUsuaScbJSzY2nev6eDoUSebx92Wi9nCib3ralD4qiYaqpgb3cERSgzWiGqlZv/ddnUXMVIooAtJQ+vb+ClC1E2NIY4M5amscrDu3a183pvnEjATVv41kqlHeyP8+I5Jzj3jrsVVtUGZrz+wNp6ckWTY0MpxlMFipZNrmjTXecnmTOYzOpkCk7v5EstTbe0VrN/fT0vnY/SFvZPFZKFAy66GkK8ZUszaxsCVAfc5Iom96+uZUdHDZPZIh5Nua3tUB9Ye3m7fYrCT89GMS2bp4+M8BuPrsHrUrFsSXd9gLDPxQvnJ9EUwf1rZrYHVRXBIxsaOD2a5u45rmeXInCpgu2LTP64E2kLX14NuW6VO2g6Uso3l3O+S3hdCu/Z1c6rPZPEs87T97Wejj98RernfFnfFJqRhy2Ewu6uCFLCyeEko6kCLlVMiXLt7a5lb3ct/15yC9m2TTTt+HiDJZVSu/Tm6Tm7A7Ec33xjEN202VCKUdzbXUd7xE/fZJZvvjHBmvoQW9uWz0l8reyk3j97YsHzel0qT2xrnvr98c1N5A2LnoksqiKoDXp45862Bc+/GKZLHEx3T1u25PmzEwzGc1gSNrdUEfCoxLMGVT6NvV21WFJStGyEIolldY4MJKbOuy2t1bNao4b9bv7ru5yW3aPJAm/d4uyTrnqnAOxtO1pv7pedB6oQuFQFTXWuXcu2+f/Ze+8oua77zvNz33uVY+cckDNAEGAAk5gkkRJpJSrLlmjJsiXbOvbYc3a8q9312LvjdZTHmrVs2atk2bSlkSxZkWIQsxhAECByajQ6d1d35fjS3T9edaEbnQq5AdTnHBx0V9WrulX96t17f+H7tSQYtoWCoC3ipWjYhL0uon43WzoiqELgmyeMt7UzOqtbfCb1QQ+fvXt1pf/nekJy5rqoqNUtIi+G0fz8g5Hyl873uRdiZ289bk3B61KXrJWuhrPlos+mLuDm3Td0MJEpsn84iWlJxtIlckUdd/DMCfbWDS20hDwcGE6xbyiFBNojXm5b1VC5yHfW+Xnn1jYyRZOXTkxi2hJNEdy1tomgR6u06D99ZIJk3uD0VJ51raFl17RzNfDWjS0cGE7TFvFe0UTozPN1ReOZXcCJ8TR7B5McHEnhd6vU+d18fFcPE2mdFU0BkBAaduHRVNyqiu0SpAoGo6k8Tx4eZ31rqLL7ne8cbo14eXibYy+50IXyUrLQ92p64js1maMz6iPo0ciVLKI+N4YlCXlduFSnV8FGEvK4UBWBbS/+vGdT53d26RvbFy4CuVaR9pnrhavKSMLFMJq/bKiKuGhdnidjWX6yf5So380jOzrnlYcwLJuX+6aIZUo0BDzkShYeTeF3v7UPVVH4/Ds3sKIpyMlYlpf74tjSJuTVkMD2nnpusv9MjQAAIABJREFUWTl7Gzu9FT81mWMwnqc14mVHz+z30xL2kswbNATcuKqcyZcTy6Efwe/WlkUz2Hzn6z88f5KfHhhD4JQPn4wV0ZQ8EZ+bh7a1sbYlxJGxNG1RH4mCQb3fRbpoInAqexqDHlRFoJs239kzxGSmxNs3z9WImS8HcakpGhbffn2IVF7nHVvmFkbkdZNXT8VpCnsca0u/G8vWaQp78boUxtJFskUDAbSGfbRFvShC0Bj0sH/IEb/rqPPxnu0di4ZKXaoy57t3vdBTfybk6apyAXnBRvNXK0dGMxiWJJYpMZEuzamTHkkWyqsvJwG1s6eO+zc0c2Q8zTNHJhECXj4VZ0VTkIMjqXLYR/AH79xAwK2xapGdyrtvaGcyq9MYnFvx88CmVm7srqM+4D4vT4PlcBGuAbmSyVRWp6PON+uC9fLJOJqiUDRM1reESIUNTsZylEyLw6NpOqI+/C6Nj97SjaYKdFPi1pzqH8OSlfMilnU0YsCpMlsOjKWKDMXz6KbN0bHMnElAVZ1Gznq/m2zJ5O0bWxlLF1jXGiKvW6xuCmJYNooiWN0c5Fd29Vbyb88em8CWksF4nlTBqFXLLcBkzplEJeBzVXd5vxjNYmuAP8FRE63ESKSUKy/0uS8lWzoiDCfzRP1uWiOzY4f7h1I8eXgcgWNMUTBssrrJSKrIeLpEumigCMHKJmeLv60rSjwXo6POx+a2yJLVPJqqzHnNaRRFLHhfjasDw7J57NUBMkWTDW1hHpixUr9nXRM/3D+Kz62R1S00TWFDe5g6v5v1reEFjzublpCHngY/sUyJLVew03cmfrdK/1SOXMlkZ+/cHfuhkTT7h1KEfC7cQrBrVQOHx9Lc1FtPW8THiqYAybxe6eafeaGf/o511vmI1pomF2RHd5R/euU0SGgMVTdRXoyA6VeB/xP4AnAP8CiwbOMYw8kCJyeybGwP8+m75m+nj2WLDCUcqYFNHWF8Lo2j485qy5KSBza3oikKXs0JIa1vDbO+df74Y0G32DOQoCHoXvAxNa4+prIlDo6kWdkUoLNu9i5SN22yZR2eRH52md6DW9rQVIUXT05SNCwe3NzG/WUl02zJXPC4ad4cSpIrWezoqeO9N16ZhPdClEy7co5rqkJRN3nstUH8bpX37+jEsmWlyTNTsljbGkJVBCsaA7g1ZZbUi2nZ7BlIoqmC7V3RRb9jNc5gSIlPU7CkJOK9fJOAT0r5lBBCSClPA38ohHgeZ2JYVti25HtvDKObNqcmc3z8tt4FH1sybYq6xfPHYtQFPDSHPbRFfNy3oZnJbAlFCDZVkXh67nisohlUH3BfsRrtGheXH+0fZSqrs384xW+8ZdWskE/Ao3H/hhYG4nl2npXz+dH+UU6MZ+ibyNHbOHvyCC5yHDhdtNM2qZYtuWNN4yV4Z+dPV72f21Y1kMgb3LaqgW+9PsRPD4wBEPW5+dSdK/nW7kE2tIboqPPxpWdPLvhdfGMwyYsnnPJav1utTQBVki7o6GVb0alsdaY6F2MSKAohFOC4EOK3gGFgWbh4D8bzPHc8RlvEyz3rmhHC0SjRTXvRqpts0eTkRBZVETSWPQBaw17u23Du4mOe8usoQlRdt1tj+TP9d3WpAoGzwHjqyASxTIm71zWxuSNS0ah6/XSCfYMJsiWTwXgBr6ZSH3Szqik4Rw9q5nFnM/OcXY5VY7pp8+ThcSazOt0Nzop/IlN0zn1NsLYlxOff6ZgPSikX/S56Ztzm0ZZXp/dyxu/SAIGUsuo+o4sxCfwO4Ac+B/wxcC/w8YvwvIti2RLdtBeVZH65b4qJtJP4Xd/qSC98YGcXg4l8JZ4/HydjWRRFoApY3xZma2eEdefZZHPnmibCXhfNYc+yFi+7VjAsG8uWVZkBXQgPb2unL5ajq86PoghGkgXeHEpi2ZLX+uOVunwpJc8fjzGSLDCcLLCtM0p3vZ/33+SEcs6O59u2pGg6/shn0x513PPyusXalnMrkS4aFooQl3Ty2N0fr/Qz/PueYe5a08jm9gguTdAUnL0DFkLwyI4O9g+n5pVI39oZdQQBFYXexoW/qwuR1815P8NrHSnApwlMG+r8lykxPK0dVN4NfE5KmbnQ51yKkmnxr68OEs/p3L2uacGy0RWNAYYSBTyawnf3OPoy77qhY0kV0daIl1TBwK0qbGmPsOkCOg8Pj6Z57niMsNfFR27pvuQXp+uZdNHgsVcGKBo279zatqA8yMXA79ZmnUcBt8qx8QzJvDFrgSGEoKfBT7ZoUgxY+N0q27vreOboBEXDJuA5U0lm2ZJv7R5kLFXklpX13LZqbrjnfCwfB6byfH/vMFpZc+tSVda0R70k8gYFw+L+gIeehiBrWpy4f9s8jVt/8bNjnJjIclNvgt9/+7o5959vmWsip/P3z/axo6eOu9Y2nddzXK00Bz0ULEciR1Gqu9ZcjOqgnTjJ4VD59xTwq1LK1y/0uRcilTeI55zEWV8st+AksLO3nvVtYU5OZHj6iKM3fjqeW1I2tzXs48HNraiKQNOW3lIZls2JiSzNIQ8NQSd8lCoYFbP6aS+AyWxpThKxxsVjLFWsKIWenspd0kngbHK6xdqWEKbtCBzO5F3bOrhvQwsuRSCEYCCeZyxdxLQk/ZPZyiSQ081y2afklb443fX+i3K+nI6X7Spti5Fk4ZJNAkIIblvVQLZk0lnvo7vBz6/duRJFmRvSKeomJyYca9iFSlxPxrK4VeWcJz69LDh3ajJ33U0CE9kSEY9WNt65fM5iXwE+K6V8HkAIcQfOpLD1Ijz3vDSFPGzuiDCWKizZFBT0aKxrDdM3maNk2FWV023rijKWLuJ3q6xsXPpC8tRhRwjLrSn86u0r8GgK33ptkGzJJOBRaQx5aAy4aY9cPO2aWj/AXHobAqxsCpArWdzQdXk7ZVvDXja1R4hlS3PCG4oiZhmheDXFUTk1bWYW0oW9Lm7sqeOF4zGKhs23dw/x/p2dFzwRbOmIMJwo4HEtrKB7MfC6FCazOnndrEhcLBSu9bo1Htjcyu7+OG/bNLcUdrpMG+A92zvOKSQU8Gg0BN3ceh02jK1s8GNKiWnZs/wmFqOqSUAI8V7gT3ESvqL8T0opw0BmegLAufEFIcQlDQkJIc7JINzrUnnP9urL6eoDbj58c/U2k4WyPZ9h2Ri2jUsKSqZVee1fvrWn6ueqcf64NeWKaeQoili0rn8mlqSSYzo7Rv+WtU24FMErpxwZ4Gnrxwsh6nfzoXM4n88Xy6YyyVQT9nz09hU8uoAta2HG+y6c42cQ9GjnrR921SOUSiXVdFHLUlS7E/gz4GEp5eF57ntVCPH3wGM4+48PAs8IIW4EkFLumTNOIW7B6SuwgN1Syt+tchzLkvs2tPDGQJKOqLey4nvXDR2cjGWvqGXf1cK57mouRHBuObCiMcA965sXbKra2VsPwrmQLtZ5vtxoj/p468YWknlj3vd1LtzYHcWyJW5NsL71yimfXm3csrKBD+zsJJbR+fDNXfxOFcdUOwmMLzABgGMtCXP7Am7DmRTuneeY08C9UsqiEOKfhRBbpJT7qxzLOZMrmXx/7wgl0+Khre00VTlDVkvY6+ItZ8Ueu+r955XEq3H1I6Xk8YPjnJ7KcceaxnkLCxYLV7k1Zd6k8HJHN22OjGVI5nV6Gi7s/NdUhV2rrr9wzoVSMi0sCZoqyFXpn7HoJFAOAwHsFkL8G/A9oNKBIKX8rpTynnMdqJRybMavJpfYBOjUZI7x9BmdlabQ9ZUsqnF5yZTMSrLzjYHkBVWXXU2MJAsMxvMAHBhO1RZBV4DBeIHhhGO0s384VdUxS+0EHp7xcx5424zfJfBdIUQL8N+Adinlg0KIjcAu4Ckcw/nDgC6lrBwrhGgHvgnUA0Up5aGzX1gI8Wng0wDd3RcWz+yq9xPyapRM+7JWjNS4Pgm6Nbrr/Qwm8mxou35CGa0RLw1BN6m8wZoraF5zPdMe9RL1u8iVzKoNhMRMs5MFHyTE7VLKF+e7TQjxE5xqoP9NSrlNCKEBb+BMIP+XlPJj8zzf3wA/Av53QJFS3rbY6zc2Nsre3t6q3tCVIJHTK2VpDUEP2nk4gqULBqYtCXm1OSWGM+nv7+difBZFw/E1BqfmPeSdvR5wylp1pISwz7UsXc4u1mdxLXCtfxZnf8dKhkXRtPG7VXxnJaGv9c9iMYqGxUTZ2Cric9F3eL+UUi7aIVhtTuCLwI0L3NYopfyWEOIPAKSUphBiOrxzT1lH6LtSyi/MOHYbsBr4PeDzQojQYk1mvb297N69u8qhXn5298d5/vgkjSEPH7qpa9ZFfCxV5Kkj4zQEPLxtY8u8CqPDyQLfem0QgJVNAd51QwfposFPD4zhUgUPbm6rVFvs3LnznD+LsVSRp49MUB9wV8aQyhs89toAumnz8Lb2WaYn4Gznpy0Nexv85HSLtoiXe9c3n5fE9aXgfD6La5Vr7bMYTOT5m6eO41EVfv/t6zg+keWF45M0hTw8sqOTLz1zEoCAR50jBHmtfRbnwu5TU3zuX/eiWzYfuKmL/+WBDXMKc85mqZzALpwEb5MQ4j/NuCsMTE+/OSFEA+XOBCHErUAKGAXW4uQQvi+EeEpK+Wb5mBagAafsdANwH06+YeZrX7Rw0KXi+eMxTk/luWN1I7/xllV4NGXORf61/nhFumJzR3jemu+GgJuIz0W6aKAKwTdfPk3JtEgXHEXJo2MZtl1A3fvu03HG00XG00U2tYfpqne8aT95xwosW5LI6/zLKwM0hTzcv8G5yLdHfXhdKqZlM5ktkS1ZxDIldNNmKqdzy4r6Zb/lX6jq6GqvLroWsWzJE4fGSOQN7tvQzI/fHGVgyskvPHV4nPft6GJze6TyHett9NM/mWdFY7B87DiJvM5965eFbNkVYzhZJF3QsWzoKzfjLcVSOwE3ECw/buY3Pg08Uv75PwH/AawSQrwINAGPSClLlJPIQogfApuB6UlgTEq5vnzffwBPn/3CUsovA18G2LlzZ3Wtb5eRVMFgd38CgJdOTvGRW7pJ5XX+8oljlEybz923mo6ok4s4OJKiPuCmzjd/p6ZbVeiIetEyghMTWSROeCjs03BrKgPxPAdH0tyx+vwqRnobApyYyBIsO1NN41IVXCo8fnCM549NEvComJZFPG9wU289n7xjBbaUPHM0xrd3D9IYdHN6Kku6aDKUyLOhLYxpSd62qYWQt6bxXmNxBqbyvHBiks4635xO3lOxLP/4/ClyJZNM0WBtS5D/+foQiiJY1xpmMJ7n+eNnjn33DR3kdYuAR2NgKl9JxO8+nbgSb23ZkMiXyOpO2Gwkka/qmEUngbJ72LNCiK+VZaLnYxXwINAFvA+4BdDOCvHcjhM+mubN8i7jTSAspVwe1kjnQMDtdAJPZkr0lmUofnJglJdOTGJL+M7uIT53/1oyRSdBowpBoqAT8GocGE7x5lCKnno/I6kCumkzlnbUFgu6ic+tsaEtzMPb2kgXTP7l1QEAXjo5eV5j3dwRYWVTALeqVLxpZzKcKNA3mcXnUvn+3hF0S9I/mWNrZxTdtCkaFutaQ9i25MBICqN8/7QUwP7h1FVZ0ljj8vKLvsnKjnRLR4RTUzmOjGbY0VPHYDxH/1QO05K8MZBgRWOAzrIrW9G0OHQyXTl2c0eE+oC74h/dEHQ7fsW6Sfd1XpG0byBZ+XkwWazqmGpzAv9DCHH2ajwF7AY+JaX8thCiDrgf+EvgS8B/FUL8Mc5u4AUp5StCiC9KKX8bp/nsG4CPZeQ7cGgkzZtDSTa2h9naGUVKyTPHYsTK9pP9kzlWNQe5qbceTVX4yM3d5HWTo2MZfu9bezFMG92ysSVM5nT+9dWBSqLX71bZczrJy31x+idzqIrg1VNT9DQEMCwbW0o8mso7trazsilA0K2hKAJNVWgIupnK6gtqHk1mS/z8yARRv5v71jfPm3dYSlHRtiXSlgwl8uR0i2zR2ekYls19G5pxqQrhgMbalhAnJ7Js745i2s423qUo/NtrA7RHfdy55vKW39bkM64euusDjCSLNATdeF2OtEosU6J/KscDm1oIeFQMU9Ia8bF3IMELJyYRwNs3NeNzuzgwkqIj6ptTxBDwaHzi9l50065MDNcrbveZ775pXoQ+gRn04YR5Hiv//kFgHCfmP92n/07g76SU3xdC/KGU8sfAj2c+SXkCQEo5xPxNZFeUZ45NUDJsBuJ5jo9n0S2bkUQBRRG83DdFV72f0ZSzivG6VFRFEPK6+LfXBhlKOCv6O9c04Xer6KbNaKrIWBoag25sCScmMgjhSEr4y2WEh0fTBNwqd61pIqubtM/oOgYnZPORm7vJ6RaRBWz1dvfHGUoUGEoUkNKJ8W/tjLKhba4Rx77BJEfG0mzvrquUkAW9Go0hLy4VBAKPpmJYFqenCkhgJFlkc2eElrCHVMFgQ3sYj6by8Zu7saXkh2+OMpIsMpIssrEtXBHRq1FjJl31Po6Nu1nREMCtqfRNZolldBQBKxtX0Rb2kcgb3NxTxxd/fgLTctadP35znAe3tLGmOYhLVcgWzTmy7E5oc/l5LFxuToznKj8XDLuqY6qdBLZLKe+a8fsPhBDPSSnvEkL8Slk24n7gT4UQHuCq/Gt01fk5MZHFsJyJwLRtDMvGp2gVM+yWsBf3WSfb2tYgA/E8rREvv3PfWnxuhb995iQHR1JE/S6kdGwmB+I5JPCrd/SyriXMa/1xDFuSL5m8PpCgIejh5b44v7Stfdbza6pCxLfwR9pZ5+fwaAafS2HfUBJNUZjKTcyZBCxb8vOjE+Xyz4nKJHDH6iYK5UnmyGiaw2MZVjT6iWV1bNtRt4xndaYyJfweFaELuur9lYqlrjo/w4kCUb+rlhuosSAvnpgkntOJ53Q2tIfJFE3SBZ1sycORsQyWlAS9KnsGk9y5ppETsRwCeOumZjrrfQwnC9T5Xdf9an8xZnoIuNSLayrTJITollIOAAghuoHpIPBJ4HHgL6SUSSFEG/Cfqx30cuKdW9pIFQzG00V+enAMn0tlbUuITMEk6td48vAEzSF3Jdzy6qk4p6dyPLi5jbdtbKUp5MHv1jAsG5eqsKopWFaKBAR0N/hRhCBTsIj63XTV+zk4ksaruREILCnpiJ670ujmjghd9X48msKP3hxlIJ6ftwpJVQTtEefL1Bbx8fjBMYqGxb3rm1ndtBrblnzsK69QMmximRJNQS8l0+KWFfXEcwZhn4sP39SFbkmi/jMX+12rGtjQFiLgWbzHocb1jc+lcmg0TXvEi1dTmcyUyOsWo8kCK5oCBD0aed1ifWuIBza34dEUgh4Xj+zoxq0pbGwL186xJbihq44nj0wBjrLtkSqOqXYS+D3gBSHESRwF0RXAZ4UQAeBrUsrvTj9QSjmKUx561aEogrqAm7qAm/Y6H4mszl8/dZxcyeRULIslnXJNVVFwaQp9sSyaovD88clZqqOaImgJexlPF9neHWXXykaKpsl39jj+xtNm2+tbw7RFfHg0BSGc3ULUf35a79Ohondv7yBVMIguEDp6345Ox+sgnufLz/dhWI6XsktTiHg1TMvGtG1S5fJUW0pSeZNHb+/F51bxaCrzifqe77hrXDtM59CSeZ271zbPCdkUDZuVjQFcqkJeNykYFoZpkyvvsP/6gzeQLph0Nfh56cQkQijkdIujYxm2dEZq51gVTGX1ys/TzXVLUdUkIKX8sRBiDbAeZxI4IqWcTj3/9bkNc/kzmizwd8/1YVs2E+kipi0pmk7ytmRaPHVoDE1TaQ17CPvceMvOZS1hD9mShWHZPLS1Fd2SNATcCCGI4OLR21agmzaRGavomXH+8/VSTeZ1nj0WI+Jz0RzycHg0w9bOSKWOP57Tee5YjPqAm/qAi6NjWYSQTOZ0bFvyUt8URcNxvXrHljZimRIuVfDk4QlsCULULvI1lmYgnuf7bwxTNCxsW/K+HV2z7ve4FE5N5miP+vC7NSxbYkkwbediFfG7iZTPs9aIF0UIFAHN4VqOqWpmRIBK5sXNCQDsAHrLx2wVQiCl/MY5HL+sSOR0njvuXBjvWN04qwv2268PcWQ07XjVagoBt8o7t7SSKVlkigaHRtIoiuAdW1rZtaqRp49MMJwo8NqpOGGfC69LpT7gnlM26XOri3oiny8v98Xpi+UASbZoEvS6mMiUKpPAk4fGefrIeGVcIa8Ly3aMJ4qGhSoEybxBrmTx2besJhpwY9k2bk3FMG0e3FKdTn6N65tcuaHQtCUTmRJ53eSZozFcqsLd65qYSJfQLZt0wSBTNEBKVAUsS5a9mCdJ5HXesraJlU1BPnF7L6oiCNZyAFUTy1T0PSldzJ2AEOKfcPoB9nJG8VPilHlelbzcN0VfLEdfLEdvQ2CW4mFHnZepbAkhBKu7I4R8bta3hbltdSMvHIsxkiwiBHhdTvNVa9jLcKJAU3nFIgS0VOnqczFojXg5PJrG41KpD3iI53Saw25+fnSCXMlkIJ4nkTfQFJPeBj8SR2NESqdRza0pbO2MEPG56Kj3VcpJf/u+NZiWXdsF1KiKznofN62oJ1cy2bWqkVf6pvjhmyOoQtAc8jCWKlLQLSbtEpoqWN0SYixVZH1riKFEgdfLjV4eTeGBzW0LVsPVWJiZuTpvlZGFaqfYncBGWY3a3FVCc9jLkbEMHpcy64MDCLg12qM+XKrA73Hhd2u0hH00h7xs76nn2EQWgagoRN61ton1rSHCPhe6ZWNZck489FJyQ1eUzjoffreKW1WYKldg/P2zJ9FNm/WtIbZ0RAh5Xfzyrm4MSzKZKXF4NI1pS96yromdPfUEPdqsnUptBVbjXAh5NDZ3RIhlSqxoCHB8PM1UVkcISOR17ljTgBBOyXTY5+Kbn7qFvQNJbl7RgGHbeFwKJaN6W8Qac7l9dRPf3j2EZcOOnjqqUVCq9lt+AGjlKk34zseOnjq66/343eqckrOJdImSaWNY0Bbx4nM7JaLguEL9xltWIRCzYvvTJ241tnqXgplyEC1hL4PxPOPpIraEkM/Fh27pJujRKqv8lrCXv3j/NnIlk9UzNIAOjaQ5PpHhxu66mh58jXNiKFFg/5CjYf+Lvim2d9dxYsJpjFzfGqYl7GFju7Pj9GgqpyZzJAsGsYzTjPnxXb3kdeuimz5dT7RFfHTWBSiaVlV+6lD9JNAIHBJCvMpsU5lfOvdhXn4sW/L88Rgl0+auNU2V1e5CJ9uO3joGE3lMSzKeLhHwWHx/7zCNQQ8rmwJX1CSkZNr8YN8IG9vDi1oPrmgMsKO3npJhsbOnnuaQM0kdG884gnSd0TkdyIZl88ShcWwpmcrq/Ood8/u/1qgxHyGPyr6hJKmCwcb2EKubQ3z6rpVoilJZME2HSS1b8vgB51yLZUp86s6VBDxarQfgArGlRLec4pTiRe4Y/sPzHtUy4OhYhjfKmhpBj8btSwixbWqP0BTyYFmS7+0doWhYnJ7KkcwbnIxlWdUUvGIr/nTB4MREloF4nt+8Z/WCj2sIevj0XSspGlZlArBsyU8PjGHZkvF0kU/duXLWMZoiqAu4mMrqVZtU16gxzdHxLJmiCRL2nE7y9k1tC3aPq4qgPuBiMqvXVv4XkbF0CcsGRQjGUhdRO0hK+awQogdYI6V8Ugjh54yU9LIlntN5pW8KVcBgPI9h2dy7oTptm+kL5y/v6iFdMHhjIMmx8QwRn+uKNqtMG9Y0BufPOZyMZTkymmFTe4iRZJFsyeT21Y0EPBpquQ9iMlOiYZ7jhRB8YGcXk9kSbZFzb1qrcWEspoN0Nchft0d8xHM6RcMi7F368vCBm7qIZUq01nIAF401zQE0RVA07aqdxaqtDvo1HG3/epwqoQ7g73B8AJYtzxyd4PRUnlimRNCr4VYVcsVzszMOejSCHo2WsJetnc4O4Uq6bEX9bh7Z0Tlv9ZGUkp/sH8WwJPuGkhXHJZemcM86R2f9Azs7mUiXaIvM/8XzutR5u41rXDyuVdG70XSBiFfD71LIlpYuT/RotXPtYjOeLlEXcGNatrMrq4Jql7S/iSMHnQaQUh4Hlr17w3SFTtTvIux1NEfOt9xRVWbr5VwphHA8k93a3D+dEKLy/lrC3spkVTfjPXs0la56/7yS0jVqXAgtYS9Br0bQ66ItWlvdXwlaw15n4ep10ValBE21OYGSlFKfbqgq+wgv+3LRu9c2sbopSNTvQjdtDEuS101+emCMrZ2RinzDtcQjOzoZTxdpi/jIlZzW/JBX44lD49QH3OzoqbvSQ6xxjdLTEOBP3rOV8UyRnb31lEyLl05O4VEVbl3ZMK/EeY2Ly6rmIB+8uYvRZJF33dC+9AFUPwk8K4T4XwGfEOKtwGeBH5znOC8bQohZZY6WLfnbnw9i2pLhZIFPXoPVL16XSk+Do+7j1tzUAT89MMrhUcffpz3qrcX7a1wyuhr8dJWrzt4YSLK3XJAR9bvZ2D5X2rzGxWUgnufkhCMn/XLfVFXHVBsT+C9ADNgP/DqOT8Dnz32IVxZFONr5wHXVjTjtT+BSBX5XrQSvxuVh+rwTgjlGMDUuDdMFIMAsX5LFqLY6yAb+ofzvqkUIwQdv6mI0VaTrOkpI7VrVQHvUR8TnmtXgVqPGpWRje9jxyVaVWhfwZaIx6OGjt3STK1kLOhGezaKTgBBiP4vE/qWUWxc59hbgCzhaQ7ullL87474/BN4DJID/kFL+VVWjvQj43dqiTVbXIkIIehvnE4CuUePSUqv+ufw0BD00nMMlbqmdwEMXMJbTwL1SyqIQ4p+FEFuklPtn3P97UsonL+D5a9S4bliorPRq6B+osbxZdBKQUp4++zYhxENSyh8u9cRSyrEZv5qcUR+d5k+FEAng96WUe6sZbI0aNWrUuLicT7H4H53Lg4UQW4FGKeWhGTf/jZRyB/AZ4IsLHPdpIcRuIcTuWCx2HsPfVw6GAAAgAElEQVSsUaNGjRpLcT6TQNXFvkKIeuB/AJ+cebuUMl7+//hCx0opvyyl3Cml3NnUVJ3UQ40aNWrUODfOZxL49WoeVG4o+ybwn88KDSGECJf/b+Tc3M1q1KhRo8ZFZKnqoPcucHsnwEyD+Xl4P3ATTuwf4A+Aj0gpfxv4cyHEZpxJ6L+cx7hr1KhRo8ZFYKlV+MOL3CeBBScBKeVjwGNn3fyL8n1V7SbOl8F4nsF4ns2dkTkNE6OpAn2xHBvawtRfRveva5lcyWTfUJK2iI8VS5SiDkzlGUrk2dLpOJ3VuDQcGUuTyhvc0B3Fc5bNYLZk8uZgko46X6W7/Foir5v8w3MneWBTW6V7ucbCLFUd9OjlGsjFomhYfO+N4Yo0xPt3dlXus23Jd/cMo5s2J2NZfmVX75Ub6DXEk4fH6YvlUITg0Tt6F+xUzJVMvrd3GMuWjKaKvG9H52Ue6fXBSLLAT/Y7Edi8bnHP+tlaj08cGqN/Mo96WvDJO1Zcc0Yu4+kSTx6e4PBohr/64A1XejjLnqr/+kKIdwKbgErrn5TynCqFLgdC4AhV2XKO7r8QziQxkizM28Z+YjzDs8di7FrdwMa22e5hqYLBoZE0PQ3+a1J47lyRUnJwJI1h2ViWzVAiT9TvQhUL1w2oiiBTNIjndNqjXvYNJhECtnREEIscV+Pc0BRR0fW/oSuKZTvS4m5VYXNHBMuWDCXy1AfcFfORU5M5NrSFzltldzlhWDZHxzKEPMve8mRZUK2fwN8BfuAe4B+BR4BXL+G4zhkpJfuHUwgE61tDvNw3xfbuaOX+v33mONmiRbromLAXSnO1tv/88aMkCwYvnJjkq4/ePOu+n+wfZTRVZM9Agl+7c+W8Us7XGgdHUuhlc4qDI2lawh7cmsLpqTxuVeHZY07pbl43OTWZoz3iZTiRJ5E32NgenhPusaUkVzJJ5HX6JrKMJB3nI0UINlfph1pjaSwpebVvkkTBYGd3hD0DCV44PgmA16VQ0i0SOQO3pmBYNt/ZM4Ru2pyIZfnlW3s4MZEhmTfY2hm9Ks/zkmmTLBi8enrySg/lqqDancBtUsqtQog3pZT/VQjxlyySD7icZEsmh0bS5EoGewdTFA2Tl/vieF0qf/mzYzSF3GSLBj8/GsO2JSBQFYFu2tS/cpqOqI/OOh+pgklONxlLFeZV2dQtm+FkgeaQh+th0XpiIsvPDo4D8IO9wzx/YorWsIcbuutQhGDmJusXJyY5NZXn1GQWv0fDlnDDRJSP3doDwGA8x1df7Gdda5Bj4xlyJQuXorC+zVGV1NQzH+hgPM9Yusjm9kjFC3omA1N5xjPFqk20rySX0zwmVTD47Ddf5xN39PKL41McLStJ/vVTx/niR3cwliqiKgJVUdg9kODgaIqhpIuiblUEx9TyruAH+0YByJTMihnR1Ui6sOzV7pcF1U4ChfL/eSFEOzAFXHYdZsuWHBhO4XWprGt1rNMee+U0L56cwu9SWFnWBCoaFsmCTl8siwQKuoVuWQjAtJ2T/Y2BBIm8jgS2d0dpCnpJFgwUIQDJ3sEkfrdasWgTUAl1yGv43JrIFDk9lcejKewZiKObNn2xLKmCyUA8T9CrEfC42NoZoTnkoWTa/Gu2REG3KBrw+uk4pg2GabG9O4qU8Cc/Pkz/VI7HD0B90I0iHMOfB7e0IhCVv2W2ZPLvbzg5g5FkgZWNQVRFVCSIUwWDf39juGJOXuMMw4kCPz86wd7BJGtaziR7x9I6um5xcDSFR1Mq36FMwSCvW0zlSrx/RycD8TxrWkLkSyZCOOf4YqG9GtcO1U4CPxRCRIE/B/bgVAb94yUb1QK8cmqKr73Yj1sVfPTWHtyayuMHxxlNFVAVwV1rGlFVlZf7nIuXbkpsaWNJiaYoICW6lChCUjIlqYKBJaGoO1Z4mqLgc6vkShaPvTqAW1X4tTtX0t3gx6OptEa8uDVlWe8E4jln8lvdHDzn+K5lS/7n60OUDJvT8RwHh9NICYqQaIpACEFRt5hIlwh6VKayOgCmLRHCEaqLZYpYUnBoNF3ZSWTLoTchBJs6wtg23NjtxKoFTiivYlhUHstAPE9fzFnNqoozUTivAchz6Fi8TpBILFs6F/YZE6QEHnttgOPjGQROWLMh4GYiU8KrqQS9WllwzDF7D3o03rO9g2TeYFNN//+6oNpJ4M+klCXgO0KIH+Ikh6uzsr+IPHVonIMjKWxbYliS9W1hVEUQ8bkQQtAfLyBt54LSHPbSGPRgI9ENm6FkASklAQ9oqkp9wEV7xEfY5+bDt3TjUhVsW3JwNEVRt+ifdC5AI8kC3Q1+3rG1jaNjaTqjPo6NZ1AVwfrW5fcl+c7rQ2RLJgeGU3zi9vPfrGWKJoblTI63rqwj6PWwviXIP70yQNGwyOkWb9/UCsCN3XXsH07hdSlMZQ1s2yZXOiMV9bl7V/Naf4KbVtSxsT3CSLKIlLIySRRNEykFnXU+3rujk7FUEcu2efGEY4oxPemGvS7ee2MH4+lS7QJ1FmGvixWNAe5e18j+4VTldgEk8gaW7Wxfp3IlHtjUhmFLuup8lYv/THoaAvQ0XK6R17jSVDsJ/AK4EaA8GZSEEHumb7vUWFKyZyBBfcBN1O8uX8ydob9nezum5fzeP5UH4OO7esiVTN6yromJTIm+WI4vP3cSy4YHNrfQHvFx+5pGEnmDOr+7Utv+xkCCLR1RJjJFpHRMWKa9UoMejR099ewfSvHkYefiNTOUsVyQZeXv84lYqYrgkR2dDEzluXdDM3+c1ymaNp/YtZLb1zaRKuh84xVHU9DrUnhgcyuWLfn4bT385MAYHREfX3jqGKm8wbbOCG/f1IotJZvaw9y3sbXyOm0RHwdmXKhePD6FaUvcmsIn71hBR9SHlBK/2zHImA7JgSNNXJMnnktTyMPHb+vlnvXNfOvVQV46mQCgOeTivvUt9E/mUBTBXWubyRQN7ljdhFaze6zB0h3DrUAHEBRCfBFoLx/jAlZe+uE5JPMGzx6N4XEJfmVXD2Gfi01tESYyRVY3BemP52mPeFnT4myDN7WHK+GFrvoAEZ+bm3obsKTk7nXN3LzCWeacvU5+9w0dHJ/I0tPgZyRZIODR5jTT2DMSAvYyTA6898ZO+mI51jSfn2dCc8hLc8hLXyzL2pYwlpS4ywlan0vjnnUtDCfy3Lqyng1tZ1bjH73FSQLXBVzsG0rx9k2tpAoG9gIf0cyV/OHRNEMJZ6c2jbiEFUOLJWzPVZr5ciZ/FyNTNBlPl3jmaIx7NjTzzVdPUzIs3rWtk1++tYeQV8Pv1rh3fTN53eTgSJq2iBe/+9rqEQDHQVABOutqRjbVsNQZ8HbgE8Aq4JeATPn2PPDlSzessyhfGwQKH9jZVbnAdzf4+c7rQwzE87g1ZcHSzdXNQT52aw+6ZS9aVVIXcHPzinoAWhZwQnJq2p2yxvXLbBcAjrNQ4zxb/HPFls7qciZuTeG37lnN6akcGxYIx2zrqmNbVx1HxzL8eL9TZWJacs4FfeZFvrcxwKGRNJ11Pryu5VnbvVwu9gsxPX1atqQ17GXXykayJYObVtWjaQqPzGia9Ls1buqtvzIDvQx4NJWuel+t7LhKluoY/jrwdSHEKSnlFXNlj/pd7FrVwMqmwJymoulYp5Ry0ZX5xTK5VhTB1s7o0g+8ylndHORtm1oomTbbZrzf1oiX1sjSKyzTtis/WwttB8oEPVpl8q1xftT53exa1cCG1jB5w6yEODXl6qvzv1ACHo2dvfW1SaBKhKwipCGE+AKwEbCllA8KITYCu6SU/9+lHiBAY2Oj7O3tvRwvVTVSwmS2hC0lXpd62Yzr+/v7uVKfhQQmM8579mjKFe8uvZKfxeUmkdPRLRtVEfPu9K6nz2IprufPwrQlU1knLB70aBw5sE9KKRddCSyVE8jgfPf9lGWnhRBpnKIDP3BZJoHe3l527959OV6qanIlk398/hS2lHTW+WZpFF1Kdu7cecU+i5Jp8ffP9mHZkuawp5IHuFJcyc/icvO1F0+RyBu4VMFn7l5dafCa5nr6LJbiev4sBqbyfGfPEAA3dEe5d33LnqWOWSocFAIQQrwmpbxJCPGGlHJ7+bbr2hIy4NF459ZWBhMFtnddWHjItGx+8OYIkxmdt25sWRam8NNib7pp89DWdppCHjyaysPb2umfyrG1ttW+rLxjaxsHh9OsagpyfCLDs0djdNX7eXBzayVEejET3jWuTloiHixbEs/r9FapoFptacBqIUQD5fyTEOJW4MouA5cBq5tDrG6+8OTweKZE/6RT3rpvKLksJoFTkzkm0s628shYmqaQ4+62ojGwpFx0jYtPc8hL83onF/PYqwPkdYujYxl2rWygriaJXqPMYNxpnG0Kejg2nq3qmEVjRUIIb/niHwN+BKwSQrwC/AuQuNABXyvEszpff6mfV/qm5txXMi1298fpiy38B2kKemgKeS57A9qRsTRvDCTmTdx21fsJeTU8LoVVTfOXmx4YTrFvMEmuaPLqqTiD8fylHvJ1gWVL3hhIcHg0jZSOzMObQ8lKCe2GtjBCgEdTODaeoWhYSzxjjeuF9qiX4+MZXjoxSUekuirBpXYCvw78Dk5/QAPgBjYDKeDPLmCs1xT//enjHBlN8/jBMb7QGJhVXvr8sUlH3VTAx27tmTep59YUPnZrD7YtHRnsy0D/ZK6iOV8ybW5dObtFNOJz8ck7nIKw+WSej45leOKQ0zT37LEYli1RFcGjt/fWzGIukN39cV466SwoBhN5Dg6nAac5cUtnhBu6onRGfXzzldO8dHKKZMG4ksOtsYx47tgk+4aSSAnfeHmgqmMW3QlIKf97uTT0IPDvwDuAh4AfArcvdqwQol0IsUcIUSz7Dc+872tCiFeEEM8IIT5S1UgvE7ppz/pdSlmRTzgbKSW6aSM40zg283KpmzZCOCs7KSXSPvNc872OdVallmnZS5ZXTmNYdlkldeH7Z1aCKTMu7IoQs+6fHlvRsMkVnQtMXjex7TNjnzlXKTN0f2q+AOfO2eeYogjyuoFhWJXSZ1s6+kymaVPUTVRFIMpnm1L7zGuUcalgmhaGaaFUqbBVrcdwPfBjoK78+0+Bv1riuePAfTiTx3x8VEp5oqpRXiaePDTO/uEU61tDPLiljZJp8a3XBpnK6dy/oWVW3bEjtjbISLJIZ52XwUSOtc2himXl9/cO0xfLUR9wMZUrEfW7Hd12S1LndzGZ1dnQFuaBza0UDYtv7R4kntN528ZWNraHGU0V+O6eYVRF8IGdXYtaYR4fz/Dj/WOEfRofuql7jgTzkbE0jx8YJ+LT+NDN3XhdKt0Nfh7e1k5Bt0BI/t+fn6Ah4MbnVhmMF2gIuPj6L05jWjb3rGvmyHiG5pCHW1fWM57WuWVFPe/Y0oZp26xqDHBoLENr2EvwGnOputSUTIt/e8352791Ywub2iM8d3Scr790Gikl90w2I4GOqB9b2vzGP79OQbf4zN2reO+NHUxmS2xqryXpaziMpwqMZ52F2+sDc8PT87FUJ8nD5X8q8DXgo+V/XwXmXx6XkVIWpZQL5Q0k8A0hxA+EEPMmmIUQnxZC7BZC7I7FYksM8+JwdNxpiD42nkVKSTynM5nVkRKOT2RmPTZbNCumKIdGMmxuj+LWVFIFoyy/7AjQ7R1M0hzyki0ajKdLWLbk9dPOx3K8/HpTOZ2ps17nVCyHbtoUdIuBJWLtJyay2FKSzBuMp+fq+h0fd+5P5I1ZEsyrm4Ns6YzQF8shpWPLd2jECT08fWSComFh2pJnj8eQEkaTRQ6UQxNHxzOsaw2xqT2C161xY3ddzXHtPJjKnvnbn5hw8kZPHokhAUvCkdEMed0i4nPx3LFJMkUT05a8eipOV72f7d11V6XxS41Lw9dePF35+dRkdTm6pcJBj5Z9hqNAENhR/hcCGoUQ+4UQb57HWH9PSnkb8KfAXy7w2l+WUu6UUu5samo6j5dYnFzJ5Okj47wxcGaeunVlPSGvxi0r6x0l0pCXNS1BIj4XN3bXzTo+7NPY3BEh7HPxjq1thH0u1rYE+fc9w/zD831sbg8T8mq8Y0sbUb+L7d11FXP7B7e0EfJqbO+O8vOjEwzG82iKYDxdpDHo5olD4/jdKk0hD20RL7YtefLQOImcPu972dYVpc7vqEh21M29EG/vjhL1u1jZFKCt3O0rpeSVvimeOTpBT72f8XQRn0tlbUuQ0VSBBze3lsNEkoe2tKGbNh11Pta3hRhNFVh9ntpE1ztjqSJPHBrnVFmltiXsZXWzc45t76rjZwfHaIt48WgKfpfKykY/ed1k72CCiFejIeimIejmgS2tS7xSjeuR//bezZWft3RWV2RS7d59BHjbjN8F8DhOfuCckVLGy/+/IIT4f87nOS6UF05MVla9LWEv7VEfO3rq2dFzRr5AVQQPbW2f93ghBG/d2DLrth/uG+GJssKo363yqTsX19h75ugEeweS5EomumXTEvbyH3tHaY14URXBb7xlFUXT4isvnEJKFkwAtkd9i8pGd9b5efSs+0/GspXkoy0lLWEvRdPi1KRBW8THnoFkZWU/mChw19omLNtmOFGkLeKjfyrPnWsWfXs15uEnB0ZJ5g2OjKb5zN2r0FSFh7c559hEushXXnT+1m/b2MLKpiD7BpNMTRWIZw2eOTbJDV3RynE1apzNN35xJhl8eqqwyCPPUO2Z9ASOYNw9wN3Al4CfSSlPSylPL3bgfAghwuX/1wHJcz3+YhAqx641ReCfx8ZwcCrPZLbIRLrAntMJLMvilVNTpArzr8bBEVyzpcS0bKJ+F7FMCduWFdPvvG6SzJ85PuR1UTQshKAyhqjfqazxu1U0ReBWFTyac1818XbdtDkymsY8K/GcKI9hmoBHw7BsioZFfcDNaKqAblggJQeGk0R8auW9NJeF5DwulXC58ifsrcX+qyGVN8jN8LOerpzSTYujo2mOjWcYTuQ4GcsSy5YQOEl8n1tFVQSKIlCFsyDxu1X8Hm1Ot3CNGtOsnNHDU+13tKpHSSl/q5wkvrN805ellAslfAEQQriAnwDbgMeFEH8E3CGl/L+BfxZC1OHkBj5T1UgvMrtWNdAa8RL2ueZo4Dx+YIyvvnQK05LEMkVMW6KpCqZlUx9w851fv60irzyT3sYAa1qC6IbFYDzPN18+jUdTKJWrhKaLc962yUkA+t0qJdPC59Z4/45OJwEY8TGUzNMa8aEoAq+i8pGbu4llS6xoDPD5Jd7X57+3n9NTeda2hPjjdztbwz0DCZ49GsPvVvnorT0EPRpuVUFKZ9I4MJTk+RNTeF2CRN6goNscGfNwz7pmiobF+3d0oaiCer8bTXXCVt31tYaxpTg+nuFH+0dxqQofvKmLxqCHh7e18cSBMf7sZ8f4+i9OE/Jq6KZdWQQoQtAa8ZLOG7yeThD1u3hoWxu3rWoshwd9tQqsGgsymj6z+p+YJz84H1Uv56SU3+UczOWllAZw/1k3P1u+7+Fqn+dSIYSoeBKfzZFxx1YxXTTIFE18bpVYpkSd30U8pxMv6LS658beJ9IlmoJeLFtyeipPZ52fk7EsnXV+prI6mioIe12MJotsao8wmioQ8TkTkC1lpfv47C7kiN9FxL907b1t2wwlnJNgZjJ5tJzAzusWybxO0KMRy5ZwawpuTeHV/jguVVA0nES0IgTxnF5xnRrPlLhjTWPl+Wp9ANUxmipWJtpYpkRj0JHeiOd1LNvGsCSZooGUjhmQQODWFFRFIatbFHSLlrCXoMfFrlWNS79gjeueZ4+eKaLJG9WVly9VIvqClPKOGUJylbsAKaW8Jj3+Htzcyqt9cVY0+NnUFmY4VeTOVX5e6Y+zoT3MPzx/Ck2B37p3zawL4orGAEXDomTZPLy1jYmszp1rGzk9mWdbVxTLssnqZkXLfUd3PemCScCjsaLxwhOtiuKsOF84Mcn968/kK25ZWe+EfYJuXj8d54tPn+CedU0Ylk22ZPKhm7r42ov9rGwKognJm8MZHtzcQsm0KBgWa1tqSeDzYXt3lERex+tSKyY/fbEsiiII+1wowKrmAImCibRlpZEw4nOVO7ZdeF0KhmXz4/2j3Lu+edn6LdRYHnzpYzt4z5d+AcCtK6JUE6tfSkDujvL/y8895RIynCyys3yhfseWNta1hvj6S/20RH3sHUzw5mASRRH8eP8oH7ypu3LcqckcXpeK16Xi1lQ+UFYW3dkzv1Z+xO/i3ds7LurY33VDB++6YfZzNgY9vG9HJ7pp8/GvvIotJV+dzHHrygbq/G6GkwUeKicnP3RzF20RHwdHUvzs4DgeTeXYeJbmBUx2aixMyOua87d4+sgE/ZM5wl4X929oobchUNm1dUS9DJd3bdMqrdOhvFQhQ1PIc02bwdS4cH6wbxS36oQLY9nqOsmrSgwLIf5ICHG/EOKqDwTndZOTseycjt2ZtEW8ZIpOvb9LFZyazNEadlZpvQ0BVEWgKmJO2KYp5EFTBEJASxXGK5eaWKbEYDyPbdu8cDzGcCJfMYRZ0RjApTpjnX4ffvcZX4Sm4Jn30lqlBsn1hG7anIxlyevm0g8uM5YqoikC3bIJeFQU4fwdpvMB3fV+6sphv+lS3uayppQiBC2hK39O1Vje3Lqy3ukxseUsb+7FqDYn0A98BPhiOTT0PPCclPL75zPQK4WUkn97bZBk3qCr3s8jOzrnfVzRsJ1EsG3z7d1DuDWF7d1RPnprN1Gfm/FUEUV1ujhn0hrx8onbe7FsecUNVyYyRR57ZbDcRKZzZCyDSxX8Hw9txLQla5tDFC0bw7SpC7iZzJYIerRKuKE5vHzey3LkB/tGGIjnifhcPHp775LJ2r5Ylu/vHeHQSIqwz8WO7no+fHMX9UEPG9rC5HSTxqCH7T11pAoGTeXQUGedn0/c3ouUXDbjohpXL5GAm6jPjWlZdFbZvFltddBXgK+Ujec/APw+8GmcprF5EUK042gMbQSCUkpzxn2bgb/DyS18Rkp5Pg1n54wtnU5fgNQiolvpgoHPpZLXTbIlk3rNTaZo0lxeiXUtotO9XJKm2aJZ0TOayDghBsOSFAyLbV1O41tQU6C8yJ9P2G65vJflSLqsqZQtmVi2RFMXnwQy5fNu2h3MkpJw+aLuc6sVqQ+PptIcmh33D9f+DjWqZCRRwK0J3JrGeOYiVgcJIf4R52I+jrMLeARYyrFmMe2gPwY+jCM98bfAu6oa7TlyajLHs0cnCHpcvNo/Rcm0ed/2DvKGzdbOhfVWVjUFeOLQGCGvi7dtbCGRN4j63XzlhVO0Rb0k8waKgIe2thPwaPTFsjx7LEZH1MdbN7Zc0hK+vG7ylRdOsbUzUslbjKYKPHFonIjPRXPIw+HRDGtaAhwaSZEqmHzqrhXsG0zRWecjr9t85YVTbOmIMBDPUzAsHtzcWqkEqlEdD2xuZd9gijUtQTRVQTdt/uzxI4wkC3zitl5eOjnFkdE0772xk7DPxYHhFHUBF+/Z3oEQgtFkgb95+gQuVbChNcx9G5qvy9LPhYxwaiY458fm9jBjqSKWhHvWVldEUG04qAFHPyiJc3GfnLmynw8pZREoLnBi10spBwGEEJdM/Wp3f5xE3uAXJ6eIZUu4VIX9I2l+857/v73zDJOjuhL2e6q7J/TkpDiaURZIAglpBEhgWQRjnGAJxvaCbQzO9uLwrb322msb7+IcwN41Ng4Ig41NsL2AFzBZSAQFQEISKMeRNNLk2LHO9+NWj1qjCT0zPUHSfZ+nn666VXXvqdvVdW4459zpvV63taaV4hzzUszLDrBoSgnLV+2iqSPKpgNNjM3PIuBz2FLTwoKKItbubqCxPUpje5SFlUVD+kJtDcdp6jD3lFACr+1tpK41Qm1LmNf2NpKfHeAv66ppDsUQgQ37mvjS22cRd5WfPbUNgMc2Hep0PttQ3cQFs8YMmcwnI+MLshlfcLS7velAE6/vbwLgT6v3Ud1oTHUfXn+AaWNyUYWsgI/rl5SztaaFnUfa2FTdRE6mj1hcWVhZZBeHsQya7z76JnHPjvPhDYdSuialiWFVvUJVz8GsIVAIPCMi+wcm5nHlditDOgLITR+TiwjMnpBPbpafgE9YNLmoz+umlObgd4ScTF/nHz0xeXr6+HyCGT4yAw6Tisyw0DTP/G9MfuaQj9tmecHCkmP3TBuTiyNCQTCDM7wezrlTS8j1vEsT9+xzhKllZm5//qRCcjKNV/JUu1LYoJlamkNJbgYicN6MEsq9GE5nVRR1LsqTMBMdV2CirZbmZVIUzGBcQRZ51gPbkgauWVSOT0AEFlb2/a6D1IeD3o3xFl6KCSf9NGZYaKC4PWx3oqp3YEJVUFVVlZrXQxfOqihi9oR8Ao5DJBYn5kJuCn+2yaU5fOKt0zqtgADOn1FK1eQiMv0OMVcR6IzfsrCyiDkT8sn0O0Pepc/PDvCpZdOOsRefOTaPypIgAcfBcYRQNE5WwMcVC4xZaPI9XzZvAuGYS1bAR9xV4q7aKJRpoCCYwc/edxYdsTh5WQGumD+RllCMAm9SPfGbgBnjv+H8KcRds07AcDw3llODS+dO4KWvltDYHmbGuHzuvKHva1JtfrwDWAHcpqoHBiFjgnoRKccogKY05Ncjibg7WRn9a2l192JM/IkD3UwCDqcTT3dlJe4z+XjCIzgZEek8nqzkLIPH73fI8+rbcZxOBQDH/2a27i1DRVl+JmX5qQ9Jp2od9JnEtoi8W1Uf6euaPmIHfRP4E8Y66DM952KxWCyWoWQgA5Hfxph+9kofsYM2AOcPoGyLxWKxpJGBDAbbPqzFYrGcJAykJ/CJtEthsViGFGuPb+mJVBea75peDp3hpU8IQtE4rirBfk4QW/pH3FXaIrFR5eXaGo6R6XcI2NW4Thh6UlqQmuJqDkVH1TM4munrjdhb3H+lH+sLjCR1rSE7JhIAACAASURBVGH+tGYfcVe5bN4EJlu7+CEh7ip/WrOXw81hFlYWsXRm+teG7i8bq5t48o0acjP9/PM5FbYRcApQ3xbht8/vYn5FoXWCTIG+Qkl/ZDCZi8hPgSrgFVX9XFL6cuB0oAOzStkfB1NOXxxsCnVGDd3X0G6VwBDREY1zuDkMwJ6kRW1Gkj117aia2D11rRGCxVYJnOxE4+a/vrdudDyDo52U/xEi8i5gDtAZz1ZVv93L+QuAHFV9i4jcLiKLVHVN0inXqur2gQjdX2aMze0MHx0M+Hlo/QHmlxdS0UsgOEvvxOIuz2+rJRp3WTqzjKyAj9xMP+dMKWZXnVmrYDSwaHKRt0JclFf2NhDwOZ3htC3pZbBDOOki4HM41Bzi7Cl27YVUSHU9gV8C7wP+BWMd9F6gso/LFgNPettPAucmHVPg9yLysIj0lc+gyfT7uHz+RK5cUM7K7bXsONzKPzanFlfD0j1vHGzhtX2NbDrQzGv7GjvTl0wv5dpzKjtDJYw0Y/KzuGzeBNrCcXYeaePpNw+PtEiWISYadxmXn8XOI60jLcoJQaozZUtU9UNAg6rejHnBT+rjmkKg2dtuwoSbSPD/VHUJ8H3gx91dnI7YQck0tEWoawuTHXCobmgnv5vwEU0dER7ZUE1da/i4Y3FX2VffTigaH7QsJwNFOQEcL9RBcdDP2t31PS5s7brK/oZ22iMx2iMx9je047rK1poWtte09KvctnCM6sYOVLuPJNLUHqXGk2N7TQtPbDpEXWsYwUwWFucECMfi7Ktv7xw2sJxcxF3lyc2HuvXstxxPqsNBiSXs2711AuqAKX1c0wgk1iDO9/YBUNV673uliHyvu4vTETsowcGmDu5bsx9XlXV76mkNxczyj4uOPe+jd63lQGMHZblZ/O2z5x1z7LGNh9ha00JhMMCHFk8+5V3+y4uCfHBxJXFX+ePqPbywvY5gho9b3zf/mHAJYJZUfL26ieyAD0TpiLg4Dry4ow6Amy6cwZLpfS+k3hGJc/dLe+iIxLsNjlXbGubel/cSc5U5E/L4zqNv0hqKMTYvi5LcDCYVBVHg/rX7OdISpqI4yFU9LCxkOXGpa4uw6WAL//XIG7z/7CEfaDjhSbUn8IiIFAI/xKwjsBsT9qE3XsSsJwDGc/ilxAERyfe+Z5GkHIaKhrYoripx16W+LUpmwMeRbhZcqGuNmPPbw8Tjx7b469pM76CpI2pbkB7FORmU5WVywFsXtz0Sp74tctx5ibTGjgiN7WYxlt21baiCqpmsT4X2SIyOiPldEr9HMo3tUWKuaS+8WdNCNObiukpzKEo45pLhd6hvjdDgydNdHpaTh1AvS8hajpJqT+AHqhoGHhSRRzCTw70uW6Oqr4hISESeB9YDe0Xka17soD+ISBFmbuBT/RVaVXlmy2EONYWZXBpk55E2ppblsGSaaU2u2l7Lrto2lkwrYWpZLpNLg4RjcTqicWaNzeG1fc0s7mbi8lPLpvHIhoNcfPoYfL5jA35NL8tlU/UBzqooHNZgcf3l1b0NbDrQzPxJhcydOPClGsKxOI9tPEQk5nLJnHG9hsh+19xx/HLFTmaOzePJN2p4/cltXLlgIss887xls8p4eVc9k4qDPPjKPtbubuD9VZPYXd+OI/CuMyakJFNJbiZLZ5ZysCnE4qkl3NLl+Lj8TLbUtLC9phUwETqLczN522ljCMdd2sJxZozJoygnwJuHWnpdWMhy4pLoo4/JtX4CqZCqEngRWADgKYOwiLySSOuJZLNQj1u89N78D/rkSEuY9ftM8NF1e+opLwpypCXMvPJCXFVW76oH4IUddUwty2XnkTYy/T78jsPLO+sZk5/J69XHBy99b9Uk3lvV/VTH1poWxhVkcbApdExY4NGEqrJiay2uKiu2HRmUEthW08rOI20AbNjfyFtm9Gzzf6glzKLJxbSGYjy68RABn8ODr+zvVAJj8rN4z7wJ7KtrZ+uhVnyOsHJHHbe+f36/5VpY2bPFx0s762hoi3CoqQMFHIGlM8qoLM3hYFOIvKwAWw+3cN25lcxIcRFuS3rozXIo3STGjg+39ryErOUofXkMjwMmAtkichZHlWw+MOz2lfF4nD+u3kdelp/CYIDG9ijlRVms3VXHmZMKWbO7DsHE03+9upH3Lapg84FmMnwOq3fV0RaOU5qTwe76NpZOL+HRjQeZWJhNaW6mif2e7Wf17noWTirkjUMtFAYzOlfvqizNoWFvI+MLstjX0E7AcUadv4GIUFkSZFdtG5NLBiZbU3uU/Y3tFOdksPNIK22RGO+cO5b/23CQyWVB9te18/DrB/jcRTMI+HzEXCU7w+Fvr+xn1rhcghk+9tS2sbCykJXbjhB3lYWVxTy9pYapZUEKgn4ONIQ4bdxR66FwLM72w62ML8imOGl1rR1HWvE7QmUP9xJXZdOBJqaU5hCKuOxtaOdwS4hILE7EBb/Ayq017K9vIyvTz9i8bOZNKqA5FGVffTtTSnOs89hJjOsOairxlKGvf8DbgeuBcuAnSenNwL8PkUw98qMntvHQa9UA/Nulp3H1wnI+efc6dta1se1wKy/vqkdR9te34/c5/PK5HVwyexwv7TzCK3saUTUx9vOyAzyztZYdtR3E4srcifkUBjNYs7ueuKvc/syOzkXAv3zpaSysLOKCWWNYWFnEntp2Hll/EID3zJtwzApfo4HL5k2gNRIjL7P/LzfXVf68di9t4ThHmkOs21OPAl/720Z8joPG42w40IKryqptdXxw8WQA7nphF82hGPsbOyjJyUAceHbLEdbtMdM9OZm7aQvH8TnC/PICJhUGKS8+2oZ4bOMh01sLONx4/hQy/T42VjfxxOYac0/zJ3RrctrQFuEfm2oYk5/Ja3sbWbu7jr31HZ3L68UUqpujVDc34HOEC2aW0hKKcf/a/TR3RBmTn8m159iJw5MVOyOQGn15DN8F3CUiV6nqg8MkU480dxyddGzqiJKXFaA9EscRIeS6qCqq5s+fIdJpztkWigECosRcxe8Ioah5ROKuS9ibQGqPxMj0+wjF3E6zptbQ0S5lflaAqHv00RqN5qKOIwOOmeKqdnpWN3ZEEREEY5aZn51BWJVE4yp5cjzivXUViKmS7fhoj5gVtgBaQzFEhFhcibpKblags/4Bwt52LK4kqjccO1q3PdVzwko0FHVpj8SIuUpPbT9VJRx3CUXinXkny2CxnKqk2lxcJSK/BSao6jtEZDawWFV/CyAi5wMzVPVOESkDclV1V7qEbAvH+MWz2zlveimOCPnZAf75bDN2/+3L57D8hd2cPbmYUMxFULYcbGH9/iamlAb585q9jM3L4JypRXREXK4+awJr9zWxoKKQJ984zLQxOVy9YBId0TgXzCpj5fZaFlQUsulgMwXZGby1S+yReeWFxOKKIzB7fH534p6w+H0Ol8+fyM7aNt52eimf+/MGwrE488oLeXxzDWPzMvnQ4km8sKOeTy6dysTiILG4Mmd8Hrf835ucUV7AslllvLijjuvOrWBTdTNxhaUzSvm/jYeYMyGfiuIg+xs6mDepsLPct88Zx2v7GykvyuaJN2qobuhg6cxSlkwrwe8TTh/XfT0XBgPMGpvHztpWKoqDVJbk8MbBJnYebqO6sYOYpxEKs30sqCjmw+dNYX55Ia2RGNsPt3L6ODsvMFq8fPuDjYiaXlJVAnd6n695+1uBPwO/FZFvYuIDzfLOCQD3AOd1k08nnr/BI8BsjNKI9XRueyROOOqytaaVmy+bi5Nko39GeSE/vuboBGNLKMpvnt/FRbOz+Nur+/E5Qm1blK+8YzaLPVv0DyyGv284yJnl5kVUmpvZGUJifoWxP18yvfuJUJ8jJ7U7+qTiIJOKg/zt1WpKc80Sdat2GB+AlnCMZbPG8vV3zz3uukvmju/cvnqhUdBnVRytp89cML1ze2qXoZ2CYIC3ziyjrjXMjsPGy3NjdRPvW1TRq6wBn8PYgiy21LSQ4fexZFoJX770NP60ei9PbD7E89tqCfgcyvKy+d1Hzj6mvImF2alWieUExBFs1NgUSbWWSlX1PrxhNu+FneijXwFcBrR5xw4AqTSx6jF+BC/1dWJ2hg9HhNnj849RAN2Rm+lnalkOjghvmVmGI8L4wmzOmHhsa/K08XkEfEJZXiZjC1Jfj/NUYdGUIgqyA2QFfFw0e4ypx4Js5g+hWWVhMINJxUF8jjBnQmrlTC3NITfTTzDD16lc5kwoYFpZHrmZAXyO8LbZNpLkqYTPEUSEmaNsvm60Ij253x9zksizwFXAE6q6QETOBb6vqm8VkdWqeraIvOIdywFeVNUzUxLA5H1xbz2B0tJSnTx5cirZHUNDe4RIzCUUdYm5LmLKI5jhIxJ3yfBaCkXBDDL8Dkdawriq1LdFiCeNLwtmvDvx7fcUkQAZfh8d3ph1oi5FQDALuodiZs7CVTMX4XdMmTmZPtrC5rqY6xJzFUcEESjMDhCJu6YloxBzlYBPCPgcDh/YT9HYCYgIY/KM8jrYFPKcqJRwzEW88hJhHeKdVhKKIDiOmPvTo2P5fRHwOWbOBYycmOEYAFdNywuMLb9/mLypd+/ezUCei3RQ2xrurNcx+Vl0ROIcajKuMwG/EI0Z50TzvDj4HPObdETjuD3MXSSer55IPL9+b5F6x3uWswI+Gg9Xj1hdJNhb305Th5lDc0QozA7QEjZ/66JggLq2SKfFzpi8TGpajjrrFWYHaEy61vh5mGN+R8jwO0S9uaeK4mxaQrHOcwuyAzS0m/nCrICPhpqRr4vRwrp161RVe23spzoc9EXgIWCqiKwCyoCrvWP3icivgEIR+RhwA/DrAcrciYh8HPg4QEVFBWvXru13Hq/ubeC5rUcQ4JV9jcTjLgsqixCEktwMGtqi5GX5+cDZFWRn+Hhicw0bq5vYeKCRldvqQBVxBFXI9JvJ5KyAQ36mD5/fz6SibKJxpT0SY3dtGxl+h7jrAsKMMXk4jpDpd9hZ20a23wxduAoLK4s8xeCybk8jDlDTEqYjGmfJ1GLGFwbJz/bTEYkjAgGfj7ZwjKUzS/nAOy/gU7fez+nj87jUG4L5x6ZD3LlqN3lZftojMXbVtjG1NJeDTSHCsTiVJdls2N9MVsBHZXGQA00dlBdl83p1M3FXKc3N4FBzGL8jqLpE4hAMOHR4E6dTS7MJx6Ewy0/A73DY8wuYNS4fV5UZY3LZdKCZMfmZXFM1adi64VVVVQN6LtLBs1sO8+reRqaU5vBPZ02kJRTl3x7cQH1bhGuqynlxRz1ba1oozsmgIDtAflaA1kiMNw82c7CxnaZQ3DQifEJUzR8xLxigNRQjHHU7LVuy/EIkrogIY/MzEYSiYICpZTkoMKUkh0vmjOP6yy8asbpI8NyWGm5YvhYXOG9aCdeeU8n3Hn0Dn+Pwo/fO4ydPbGX1rjqCGX6e/OL5VN3yDGCU2x9uPJsP/m41LrB4Sgk+B57fbsKKXL+4gqyAn4c3HKAsN4t7blzEU1uOsPNIG/MmFbB4ain3rt5LSyjG2+eO5bp3XzjidTFa8Py5ej8nxZ5AFvBZjMloC8Z57OeqGvKOvw24BPN7Pq6qT/RDyGfpoydQVVWlA/1Rw7E4Acch5iox1yU74CMcc8kK+DqPJQ8xJRzBquvbyc0KEACOtEeoLM1h++FmKouyaY8pPiCY6Tf+BcEMDjZ2UJYboDUcI6ZQHMzoPFbXGiYvww+OEIm55Gb5O8sxcYwccJWWSIyS3ExC0TiZfiOzYLq3CZmrqqpY+eLLxzmrtYZiZPkd/H6Hw80djMnPpiMSJxSJU5Rr5CvKChAICHVtEcbkZ9PUHqU1FGVicZBth1oYn5dJwK+8UdPG/IpiDjWGaA2FmD6usPMefD46r4/FTUs34HMIx+Jk+BxEhi+m0kgqAaDzd0rccyzm0hEzVlGu69ISipGd4cfnCNG4a5Qs0NAWxi9CR9Ql4BdyfA7NUfNcBFCOtEcIRVwKg5koLo4oMReyA37vuQC/Twg4QkyVTL9vxOsiQXNrmJrWMDO8yfymjgg+IDfb+H8k/kOBQICWlhae3NrAFQsrur12495aApl+Zo01c3d1rWEKs/34fD5j7eX9J8CYN0ddt7Muai++uUcZT6YJ5L4myUVknapW9ZZHqj2B32N8A77j7X8AuBsTUhrvpZ/yi384yfSbhyTDETK8KZDEg5M4lkzi2MQkO/ZKb9hj+hjzcBYkWWAmgqWN9yYaC/3+446V5B6dc8jwHytDblI005IM3zHHkqMgJr/0u/NWTs5nTL6RJTvDR7aX5/ikidDE8YJggALv3mYkWcrMrzDyjivMIrF8RPI9JK73J7X4u6vLk52uv4Pf75Dn/b6O4xwTSM/nHD03UX/JJLvDTcrKOO54T4y2Ws/PzSQ/6VkpyD72XhL/IYC8vDyuWJjX47VzK44NKpj8DIrIMfXvOEKmM9pq48QgVSUwS1XnJe0/IyJxEWnu6QJV7dV+UkQCwKPAPOBxEfl3VX05RXksFovFkgZSVQKvisi5qvoSgIicA/xKVT8tIt8GDmF6BgJcSwrWQaoaxUQXtVgsFssIkaoSOAf4kIjs9fYrgDdE5HVgmqomxxG6XUReBn6QRjktFovFMgSkqgQu7eXYX0TkWsz6AoqZLxh98RQsFovFchwpKQFV3dPTMRG5CrjN+wCsBP558KJZLBaLZagZdBxdVd0NXD54USwWi8Uy3Azaq0dEykXkryJyWERqRORBEbELt1osFssJQDpcO+/EeBNPwCxA87CXZrFYLJZRTjqUQJmq3qmqMe+zHBNWwmKxWCyjnHQogVoRuU5EfN7nOqAuDflaLBaLZYhJhxK4AbgG4zB2EBNY7oY05GuxWCyWISYd1kF7MesJWCwWi+UEIx3WQXeJSGHSfpGI/G6w+VosFotl6EnHcNCZqtqY2FHVBuCsNORrsVgsliEmHUrAEZGixI6IFJOGYSaLxWKxDD3peFn/GHhBRB7AxA66BrglDflaLBaLZYhJx8Tw70VkLXAhJpT0laq6OXFcRIq8ISKLxWKxjDLSMmzjvfQ393D4KWBBOsqxWCwWS3oZjhXBh2/RWYvFYrH0i+FQAn2vZG+xWCyWEWE4lIDFYrFYRil2OMhisVhOYdLhMTxNRDK97WUiclOyBzFw0WDLsFgsFsvQkI6ewINAXESmA78FpgB/TBxU1fo0lGGxWCyWISAdSsBV1RhwBXCrqn4BGJ+GfC0Wi8UyxKRDCURF5APAh4FHvLRAGvK1WCwWyxCTDiXwEWAxcIuq7hKRKcA9acjXYrFYLENMOsJGbAZuAhMiAshT1e8NNl+LxWKxDD3psA56VkTyveih64E7ReQngxfNYrFYLENNOoaDClS1GbgSuFNVFwIXpyFfi8VisQwx6VACfhEZjwkh/UhfJ1ssFotl9JAOJfBt4HFgu6quEZGpwLY05GuxWCyWISYdE8P3A/cn7e8ErhpsvhaLxWIZegatBEQkC7gRmANkJdJV9YbB5m2xWCyWoSUdw0F3A+OAtwPPAeVASxrytVgsFssQkw4lMF1V/wNoU9W7gHcBZ/R1kYj8VESeF5Hb0iCDxWKxWAZAOpaXjHrfjSIyFzgETO7tAhFZAOSo6ltE5HYRWaSqa3o6P+4q6/bUM6Egi/vXVlMY9HPh6WM53BLGB9z5wm6Wzighw+/HEaGmuZ1nttRy3dkVPLq5hkWTi7l/7V5awnHOm17Ak5vquOH8yTR0xJhWmsO8SYW0hGLUNIX4/Ut7uGr+BFyBwmAGF50+FoCa5hBPvlHD/PICEMER4fTx+WmovvRS3xZhV20r08vyKAj2P3rHkZYwe+vbmDk2jy8/sJ6mjhhfe+csfvD4NqoqCynKyeCxjTXcdNE0GtpihGJxirJ8/L8HXuesSYWcMamApzYf5qvvPI261igxdVlQXsg9L+9l0ZQi6tsirNpex2eXTePNmlYcEZbOLGFLTRvlRdnEXeVgU4jZ4/PYcaQNv084bdzRet7f0M7hljCzx+cT856LuKt0RGLsawjx8Ppq9tW2UtsWxXXNikZleQHG5WczuSyXTy+bhs/n8NLOOpbOLGNSUTCNtT9y1LaEueznK/ncRdO5aPY4Jn/l7wD853tO5/J5E7lu+Rrys/zc89FzeWHbEb750CbOn17KNy+fS21rmD11bcwYm0d+VoDdtW00dkSZOyEfv88uOXKyI6qDW/hLRD6KiSR6JnAnkAt8Q1V/2cs1nwGOqOp9InIVMEFVf97T+VNOP1Nv+tkDvLDjCNUNIVSVM8sLOW18Pveu3ktHJEZcYebYXOKuyxuHWnGAuCrBgI+2cBy3m3wnFmbhKiyZWkJ5cZA7Vuwg7ioKLKgoAuCLb5vJOVNL+NL969lb30445rJ4agkZfod3njGeWePyBlx3A6Gqqoq1a9f2ePzXK3bSGo5RFAxw/XlT+pV33FXuWLGTUDTOii2HWV/dBIDPAUFwXUUcQYBMv8PsCQUArNlV31m/jrd6RIbfYV65iSje0B6hLRzDdZXWcAyAnEw/lSU5AJw2Lo/CYMYxsgR8QjRuns13nTmemWPzaA5FWb5qN3FXOW1cHp987yVc990/Ut3QwcGmEIeaOmgKxXq8vwAwqcQog7ZwnNLcDP7n2oX9qqPRStb4GUy4/lYy/T7ceJxw0gM/rTSHHbVtAFxy+hhWbq+jPRpHBO788ELeONRGKBpnTH4mF58+lntX70UVFlQW8daZZSN0RwOnqqqK2otv7vH47u+9axilGVoSyr4riXsUkXWqWtVbHumwDvqNt/kcMDXFywqBHd52E2ZS+RhE5OPAxwFKxk7wykoqt/Nbj9k/ZjHLFPSbAm4vecRds+MmCk8Swh2kAh0KjquPAV4fT7o3VRDpel5P13urCB1z/VGZEnklX58oK3GeiBz7MyZVfWLbVZOhqpcvSl8NmkT5x+Rx0qHH3VdyvcRc7fyNASKxo/uJujla3ydlBVm6MOCegIh8sbfjqtpj6IguPYErgXJV/Vkv5x8B9gxI0GMpBWrTkM9I5J/IewHwyhCXNVL0954SdTGUjOZ6TpZtsHWRzvsc6bwWAHvTKMOJSKLeKlW11+7cYHoCiXGQzsZfEn1plheBTwD3YUJMLO/t5L5uIlVEZG1fXaPRmn/XvIf6XkaC0XhPo1GmBOmU7WTLazT/bsNBf+5/wEpAVW/2CrsL+JyqNnr7RcCP+7j2FREJicjzwHpVXT1QOSwWi8UycNJhHXRmQgEAqGqDiJzV10Wq+rk0lG2xWCyWQZAO+y/Ha/0D4IWUTodyGQruOIHz75r3UN/LSDAa72k0ypQgnbKdbHmN5t9tOEj5/tNhIvoh4KvAA5i5gGswq4zdPaiMLRaLxTLkDFoJAIjIbOBCzATxU95qYxaLxWIZ5aRFCVgsFovlxMT6hFssFsspzCmhBERkroi8X0QWpTHP8d63iMg/ichXvTLSEZ77MhEJdkmbIyKndUk7Z7BljSY8J0KLpd+IyEIR+YyIfF1EPisip6yPQIJU33cn7XCQiDymqpeKyOeBi4C/A+cB1ar6lTTk/7SqXuhFQe0AngbmA1Wqes0g8z6A8ZCuAf4KLMKE2ogBJcANqnokIcNgyhopPB+RxMOXcDacA2xU1aUjI9XoR0RygCKgUVVbR1qe0YCI/BTIBJ7EhKHJxzihxlX1ppGUbTgQke4a8wI8pqpv6+v60WrKmQ4SEcmuAC5QVRf4pYisTFP+iRBdc1T1Ym/7HyLyTBry3qKqF4jIFOBK4DrgZeB/gZXA/SLypTSUM5L8FRN0cLmqPgsgIo+q6jtGSiAR+byq3ioi84CfY5SUH/iKqj4/UnJ5sl0I/AfQ7H3yRSQP+I6qPtnPvN6vqn8SkQqMY+c4oAFzn/026vCiB/8X5uXrYOqtEfimqm4YBrkWdtNw+KuIrOjPfZzAtAIvYV78yQ2rM1O6WlVPyg8mpPXvgf1AdlL62jTl/0HgN5jIqfcAH8O8OH6Yhryf6bK/CrNYz8e9/SJMz6ZmpOt5kPeZAXwa+BNwGfDoCMvztPf9D8w6GWBisKwaBXW1Egh2ScsZiGxJ9/m/wHne9izguQHK9jwwvkvaBOD54ZAL+AnwS+Bq4BLv+3bg1pH+3Ybp2VgHFHST/kQq15/MPYHEePl/YIZREJFcb3/QqOrdIvIUZkW1sZgW429UdX0asv9el/0vABFVvcMru0FELgPem4ayRgxVjQC/EJE7MEo1HXU3GIq9Fnexqm4HUNVaERkNY6ZhzGJNLyelnQGEBpBXtohMBUpVdRWAqm7pYVghVbrGD+u6P2RyqeoXvSgFi4GZmF7IHar66gBkOBF5N2ZIuisp9apP2jkBi6W/iMg3k3ZvU9VGb8jlh6r6yZGSCzoNEb6C6eI7QBzY4MlW3c+87vQ2Ffhi0n3+QVUvG4Bsc4D/xPRQE8NBdcC3VPX1kZLLkhpWCVgsFsspzClhImqxDIbRvA62iPS4DscA8vpquvJKZ37plstyLFYJJCEiy0TkkR6OPZtu22MRKRSRT6dS/kgxGJlE5NsicnE36Z15ettLko4tF5GrBy7x4OjOHwO4d0SE6YFkvxcdoAlkD3b1v02zqOmyxEtXPpZusEpgZCnEWMeclKjqN7Rv88VlwJI+zhkWROTHmGCIXxGRh0UksZjRd0ZQLMD4vXjfnwe+i3l2bhKR7w4gr58CNwIHMQs8VQMfAb4+QNmcbj4+4BsjkU9/EZFvici/DmUZg0FEhtQf5ISzDvKcZe7DmEz6MBNS2zFmYrmYJdWuV9WDIvIs8BpwNsaG+QZVXS0iZwO3AtmYWfWPqOqWfshwCXAzxkFlh3d9q4jsBu4C3oNZ1/y9qvqm9zL5I8bRaw1wKbAQYwU0TUReA57AmH3misgDwFyM6dd12sfEzUjUiXf+V1T1ShG5HGPmls+/7AAACP1JREFUWYBpWGxW1akishx4RFUfEJFLvfxr8ZZBFJHJwCeBuIhcB/yLl/1SMcuXjgO+rKoP9Hb/aaRKVd/qyXYmo8sfI51+L+m2qx+cnXr687H0h5G2cR2ATexVwK+T9guAF4Ayb/99wO+87WcT5wJLMd6oYF5+fm/7YuBBb3sZ5qXVXbnPAlUYu/EVQI6X/m/AN7zt3cC/eNufxpiMAvw38FVv+1LMA14KTE7IlFR+E+Zl7mBaaeePxjrBNCB2eds/wii384C3Avd66csxNttZwD5gBuZPfV8iT+BbwL8m5bscuN+7/9nA9mF8tlYBGUn7o8YfgzT6vZBmu3oGaaee7nxSKOdDGMuq9cDdyc8gxt9njXfsQTzfDIw59kYvfYWXNgdYjWlUbQBmpFqml1YJPOWlPwVUeOlTMP/9NZgGXWtSPl/y0jcAN6ejPk64ngDwOvAjEfk+8AjGo3Au8ISIgGkJH0w6/14AVV0hIvkiUohZH/kuEZmBeSEH+lH+uZiX0yqvvAzMD5bgL973Ooy3L8D5mBYcqvqYiDT0kv9qVd0P4PUQJtP3mOiw14mqxkRku4icjulV/ASjVHwY56FkTsMojG3efd0DfLyX7P+mpqW7WUTG9nHv6eQLmGGWwzDq/DHS5vei6berH5Sd+hDk0yOeOevXMM5otWIWwUqeV/mLqv7aO/e/MMNmP8cMSb1dVau9/wuYXuxtqvoHEcnAPPuplgmmcfh7Vb1LRG4Afgb8E3AbcLuq/l6S4ml5IxAzMP83AR4SkaWqOijP6BNOCajqVhFZCLwTMzb6BLBJVRf3dEk3+/+J8cq9whuSeLbrRSLyOMYJbK2qfjT5EKZl8oEeygt733GO1m9/HGfCSdvJefTICNbJ85g/aBQTt2U55o/Q3fhqf2yRk+tgIE5HA0K7WetaVeOYoa4RRVX3dJPWCjw6wPxeBdLiTKWqB3tIj41EPn1wIfCAqtZ6edd7DaUEc72XfyFmKPVxL30VsFxE7uNoQ+9F4GsiUo5RHttSLdNLX8zRhuLdwA+87fMwvftE+ve97Uu8T+J3y8UohUEpgRNuYlhEJgDtqnoPZhjiHKBMRBZ7xwOe5k3wPi/9fKBJVZswwyUJB5vruytHVd+uqvO7KAAwY5bnich0L9+giMzsQ+yVmBXXEto8sRxnC6YFPihGsE5WAJ8HXlTVI5g5j9OATV0ufROYIiLTvP1kBZqWOrBYUiR5vqE7lgOfVdUzMPN+WQBqnAW/DkwCXhORElX9IybcSQfwuBhv84GUmUB72E7O57vef3C+qk5X1UFbdJ1wSgDjKr/aGyr5GqabdjXwfRFZjxmfS7Y2aRCRFzBjoDd6aT8Avisiq+ihC9cT3svueuBeEdmAUQpdTQq7cjNwiYi8gmk5HwRaVLUOM6y0UUR+2B85ujBSdfIypmeQaIlsADaoN3iZQFVDmOGfv3sTmcmt2oeBK0TkNRF5S4rlWiwD5SngGhEpgc410ZPJAw6KSAC4NpEoItNU9WVV/QbGuGGSmBAXO1X1Z8BD9DyB3VOZLwDv97av5eiw76ou6QkeB27whgERkYkiMib1W++BdE64jLYP3mTuKJAjk6OTrouB1071OhkNH4wyn5DCecuBq7tJn0zSxH4a5VoGLOmr/NFYJymW8UJfeWJ6mMGkY60DKauHcj7M0Une5Rw7MfwpYJf3P/k5JsotmCGg173rbsO0yr+K6fW+BjyGiTmVUplJz8/T9D4x/BWOnRj+nCfH69450wZbHyfcnMAJSgVwn5hAWBGMBYJl5Lke88c8MMJydGUZxlzyhREo+3qGuE5UNRW/kM9jovO2D0H5d2FMubs7djvGUqpr+pXdnP5d7zOgMlV1N2a+oOu5uzCNxQTfSzp2G0YJpY0TcTgoZVR1maquHQVybFPVs1R1nqouUtU1IyjLqKiToUBEJovImyJyl4hsEJEHvDmbhSLynIisE5HHRWS8GK/kKuAP3lBUtoh8Q0TWeMNzd0iXGcM+yvaJyA+96zeIyCe89GVivM0f8GT7QyJfEXmnl7ZSRH4mIo/IUd+JL3QZIlsqIi+IyE7ph0f1cNeJiPxCjEUVIvJXEfmdt32jN+Ha6fwkhv8Wkc0i8ndgjJd+EyYU9TOStD6HiNwiIutF5CUZXquxk5qTWglYTklmYcwdz8QsvvIZTLf+alVdCPwOuEWNA9pa4Fo1k2wdwH97Snouxmnu3f0o90bMJPsizEpwHxOzKBDAWZiW7WxgKsawIAv4FfAOVT0fKIPO1uEvgZ96ciXMbcdjTI3fzfGhxkdTnawAEopronfPeLJ3NR2+wpPtDEzveAmAmjH2AxinuAu8c3OAl1R1nlfGqOtNi0iJpzy7fkpGWrbesMNBlpONferFoscMJ/w7vftMJHOBiHwZCALFmPHeh1Ms9xLgzKRWegHGfC9C974frZhJxV3e+fcydL4Tw1knzwOfF5HZwGagSEwY7MUca48Pxq/kXjVmuAdE5Ole8o1gfGDA+OD0uWzicKPG0GP+SMvRX6wSsJxsdDWta6F3nwkAvJb5LzCT5vtE5Ft45oFJ55yDab2DscBKXjpRMN7ij3e5Zhnd+3701/9hML4Tw1YnqvqQiBRhPONXYBTHNZjJzZYUZOuJqHozo6ToP2NJDTscZDnZqBDPPwLjj/ASPftMJPsoJF5utZ4J3nHj7mpMBBM22g91Ofw48CkxpoWIyEwxMZ164k1gqjcHAJ7vRjdypYPhrpMXMcNfKzA9g3/l+KEgvOPv9+ZTxgMXJB2z/iPDhFUClpONN4APi/HhKMYb+6Z7n4nlmCBsr2Fa2r/GmN79DWOe1x9+gxn+eEVENmJaxz22Vr3x9k8Dj4nxnajBxI2C9PtODHedPI8xid6OCRZYTPdK4K/ANi//24Hnko7dATyaPDFsGRrsymKWkwavVf2IN4k56hGRXDXRZwX4H2Cbqv40zWVM5gSqE8vwY3sCFsvI8TGvxb0JM5H8qz7Ot1jSju0JWCwWyymM7QlYLBbLKYxVAhaLxXIKY5WAxWKxnMJYJWCxWCynMFYJWCwWyynM/wf5OJx7K4sgVwAAAABJRU5ErkJggg==\n"
},
"metadata": {
"needs_background": "light"
}
}
],
"source": [
"# Merkmale auswählen und berechnen\n",
"''' Hilfe\n",
"Plotten Sie die Merkmale in einem zweidimensionalem Scatterplot. Unterscheiden Sie dabei „gute“ \n",
"und „schlechte“ Merkmale voneinander. Eignen sich die gewählten Merkmale für eine Klassifikation oder eher nicht?\n",
"'''\n",
"from pandas.plotting import scatter_matrix\n",
"from matplotlib import pyplot\n",
"\n",
"scatter_matrix(iris_data)\n",
"print(list(iris_data.columns))\n",
"pyplot.show()"
]
},
{
"source": [
"### Auswahl der Features\n",
"An der Scatter-Matrix kann man erkennen, dass sich die Merkmale \"petal-width\" und \"petal-length\" besonders gut für die Klassifizierung eignen, da die beiden Klassen in diesen Merkmalen im Gegensatz zu den anderen Merkmalen nur eine geringe Schnittmenge haben."
],
"cell_type": "markdown",
"metadata": {}
},
{
"cell_type": "code",
"execution_count": 24,
"metadata": {},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
" petal-length petal-width\n",
"0 1.4 0.2\n",
"1 1.4 0.2\n",
"2 1.3 0.2\n",
"3 1.5 0.2\n",
"4 1.4 0.2"
],
"text/html": "
\n\n
\n \n
\n
\n
petal-length
\n
petal-width
\n
\n \n \n
\n
0
\n
1.4
\n
0.2
\n
\n
\n
1
\n
1.4
\n
0.2
\n
\n
\n
2
\n
1.3
\n
0.2
\n
\n
\n
3
\n
1.5
\n
0.2
\n
\n
\n
4
\n
1.4
\n
0.2
\n
\n \n
\n
"
},
"metadata": {},
"execution_count": 24
}
],
"source": [
"iris_features = ['petal-length','petal-width']\n",
"X = iris_data[iris_features]\n",
"X.head()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Bevor wir nun einen Klassifikator auswählen und trainieren, sollten Sie einen Testdatensatz erstellen, beiseitelegen und nicht mehr hineinschauen.\n",
"\n",
"#### Testdatensatz erstellen\n",
"Bis jetzt haben wir kurzen Blick auf die Daten geworfen und sicherlich sollten Sie noch eine ganze Menge mehr darüber lernen, bevor Sie entscheiden, welche Algorithmen Sie verwenden. Das kann aber dazu führen, dass Ihr Gehirn (ein erstaunliches Mustererkennungssystem) anfällig für eine Überanpassung des Problems wird (engl. Overfitting). Sie auf ein scheinbar interessantes Muster stoßen, welches eine bestimmte Art von Modell bevorzugt. Dieses Problem wird auch als *Data Snooping-Bias* bezeichnet. \n",
"\n",
"
\n",
"\n",
"
\n",
"\n",
"Einen Testdatensatz zu erstellen, ist sehr einfach. Wählen Sie zufällig einige Datenpunkte aus (meist 20% des Datensatzes) und legen Sie diese beiseite. `train_test_split(...)` ist eine Funktion in Sklearn zur Aufteilung des Datensatzes in Trainingsdaten und in Testdaten. Die Funktion hat mehrere Parameter. Ein einfaches Beispiel für die Syntax würde wie folgt aussehen: \n",
"\n",
"`train_test_split(X, y, train_size=0.*, test_size=0.*, random_state=*)`\n",
"- `X, y` Als erster Parameter wird der Datensatz angegeben, den Sie verwenden möchten.\n",
"- `train_size` Dieser Parameter legt die Größe des Trainingsdatensatzes fest, die zwischen 0,1 und 1,0 liegt.\n",
"- `test_size` Dieser Parameter gibt die Größe des Testdatensatzes.\n",
"- `random_state` Der Standardmodus führt eine zufällige Aufteilung unter Verwendung von `np.random.seed(any_number)`.\n",
"\n",
"\n",
"## **Aufgabe 2.2:** \n",
"Teilen Sie den Datensatz in einen Trainings- und einen Testdatensatz auf. Verwenden Sie dabei einen 80/20 split."
]
},
{
"cell_type": "code",
"execution_count": 25,
"metadata": {},
"outputs": [],
"source": [
"from sklearn.model_selection import train_test_split\n",
"\n",
"X_train, X_test, y_train, y_test = train_test_split(X,y,train_size=0.8,test_size=0.2,random_state=np.random.seed(42))"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"#### Trainieren und Auswerten\n",
"Wir wissen nicht, welche Algorithmen für dieses Problem gut geeignet wären oder welche Konfigurationen verwendet werden sollten. Aus den Darstellungen haben wir die Idee gewonnen, dass einige der Klassen teilweise linear trennbar sind, so dass insgesamt gute Ergebnisse zu erwarten sind. Lassen Sie uns drei verschiedene Algorithmen untersuchen:\n",
"- Logistic Regression (LR),\n",
"- k-Nearest Neighbors (KNN) und eine\n",
"- Support Vector Machines (SVM).\n",
"\n",
"Dies ist eine gute Mischung aus einfachen linearen (LR) und nicht-linearen (KNN, SVM) algorithmen. [*Scikit-Learn*](http://scikit-learn.org/) enthält bereits effiziente Implementierungen der drei Machine-Learning-Algorithmen."
]
},
{
"cell_type": "code",
"execution_count": 26,
"metadata": {},
"outputs": [],
"source": [
"from sklearn.linear_model import LogisticRegression\n",
"from sklearn.metrics import accuracy_score\n",
"\n",
"model_lr = LogisticRegression(solver='liblinear', multi_class='ovr')"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Um das Training zu starten, reicht ein Aufruf von `.fit(..)`. Die Funktion nimmt als Parameter den zuvor erstelten Trainingsdatensatz entgegen."
]
},
{
"cell_type": "code",
"execution_count": 27,
"metadata": {},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"LogisticRegression(C=1.0, class_weight=None, dual=False, fit_intercept=True,\n",
" intercept_scaling=1, l1_ratio=None, max_iter=100,\n",
" multi_class='ovr', n_jobs=None, penalty='l2',\n",
" random_state=None, solver='liblinear', tol=0.0001, verbose=0,\n",
" warm_start=False)"
]
},
"metadata": {},
"execution_count": 27
}
],
"source": [
"model_lr.fit(X_train, y_train)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Nach dem Training könnten Sie das Model evaluieren. Übergeben Sie hierfür der Funktion `.predict(...)` mit Daten. Als Qualitätsmaß verwenden wir die Genauigkeit (engl. Accuracy), die den Prozentwert der korrekt klassifizierten Prädiktionen angibt:\n",
"$$ \\mathrm{Accuracy} = \\frac{\\mathrm{TP} + \\mathrm{TN}}{\\mathrm{TP} + \\mathrm{TN} + \\mathrm{FP} + \\mathrm{FN}} $$"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"
\n",
" \n",
"
"
]
},
{
"cell_type": "code",
"execution_count": 28,
"metadata": {},
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"1.0\n"
]
}
],
"source": [
"model_lr_predictions = model_lr.predict(X_train)\n",
"print(accuracy_score(y_train, model_lr_predictions))"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## **Aufgabe 2.3:** \n",
"Vervollständigen Sie die Code-Zellen für die beiden anderen Klassifikatoren und geben Sie jeweils die Accuracy für den Trainings- und Testdatensatz aus."
]
},
{
"cell_type": "code",
"execution_count": 29,
"metadata": {},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"KNeighborsClassifier(algorithm='auto', leaf_size=30, metric='minkowski',\n",
" metric_params=None, n_jobs=None, n_neighbors=5, p=2,\n",
" weights='uniform')"
]
},
"metadata": {},
"execution_count": 29
}
],
"source": [
"from sklearn.neighbors import KNeighborsClassifier\n",
"\n",
"model_knn = KNeighborsClassifier()\n",
"model_knn.fit(X_train,y_train)"
]
},
{
"cell_type": "code",
"execution_count": 30,
"metadata": {},
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"1.0\n"
]
}
],
"source": [
"model_knn_predictions = model_knn.predict(X_train)\n",
"print(accuracy_score(y_train, model_knn_predictions))"
]
},
{
"cell_type": "code",
"execution_count": 31,
"metadata": {},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"SVC(C=1.0, break_ties=False, cache_size=200, class_weight=None, coef0=0.0,\n",
" decision_function_shape='ovr', degree=3, gamma='auto', kernel='rbf',\n",
" max_iter=-1, probability=False, random_state=None, shrinking=True,\n",
" tol=0.001, verbose=False)"
]
},
"metadata": {},
"execution_count": 31
}
],
"source": [
"from sklearn.svm import SVC\n",
"\n",
"model_svc = SVC(gamma='auto')\n",
"model_svc.fit(X_train,y_train)"
]
},
{
"cell_type": "code",
"execution_count": 32,
"metadata": {},
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"1.0\n"
]
}
],
"source": [
"model_svc_predictions = model_svc.predict(X_train)\n",
"print(accuracy_score(y_train, model_svc_predictions))"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## **Aufgabe 2.5:** \n",
"Entwerfen Sie einen einfachen binären Klassifikator, der nur eine Zahl in einem Datensatz erkennt. Verwenden Sie hierfür den MNIST-Datensatz. Dieser beinhaltet eine Sammlung von 70000 Bildern handschriftlicher Ziffern, die von Oberschülern und Mitarbeitern des US Census Bureaus aufgeschrieben wurden. Jedes Bild ist mit der dargestellten Ziffer gelabelt."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"##### Vorbereitungen\n",
"In dem Datensatz gibt es 70000 Bilder, jedes davon hat 768 Merkmale. Das liegt daran, das jedes Bild aus 28 x 28 Pixeln besteht und jedes Merkmal die Intensität eines Pixels von\n",
"0 (weiß) bis 255 (schwarz) enthält. Betrachten wir eine Ziffer aus dem Datensatz. Dazu muss der Merkmalsvektor eines Datenpunkts herausgreifen werden, zu einem Array mit den Abmessungen 28 x 28 umformatieren und mit der Funktion imshow() aus Matplotlib dargestellt werden:"
]
},
{
"cell_type": "code",
"execution_count": 33,
"metadata": {},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"(60000, 785)"
]
},
"metadata": {},
"execution_count": 33
}
],
"source": [
"mnist = pd.read_csv(\"mnist_train.csv\")\n",
"mnist.shape"
]
},
{
"cell_type": "code",
"execution_count": 34,
"metadata": {},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
" label 1x1 1x2 1x3 1x4 1x5 1x6 1x7 1x8 1x9 ... 28x19 28x20 \\\n",
"0 5 0 0 0 0 0 0 0 0 0 ... 0 0 \n",
"1 0 0 0 0 0 0 0 0 0 0 ... 0 0 \n",
"2 4 0 0 0 0 0 0 0 0 0 ... 0 0 \n",
"3 1 0 0 0 0 0 0 0 0 0 ... 0 0 \n",
"4 9 0 0 0 0 0 0 0 0 0 ... 0 0 \n",
"\n",
" 28x21 28x22 28x23 28x24 28x25 28x26 28x27 28x28 \n",
"0 0 0 0 0 0 0 0 0 \n",
"1 0 0 0 0 0 0 0 0 \n",
"2 0 0 0 0 0 0 0 0 \n",
"3 0 0 0 0 0 0 0 0 \n",
"4 0 0 0 0 0 0 0 0 \n",
"\n",
"[5 rows x 785 columns]"
],
"text/html": "
\n\n
\n \n
\n
\n
label
\n
1x1
\n
1x2
\n
1x3
\n
1x4
\n
1x5
\n
1x6
\n
1x7
\n
1x8
\n
1x9
\n
...
\n
28x19
\n
28x20
\n
28x21
\n
28x22
\n
28x23
\n
28x24
\n
28x25
\n
28x26
\n
28x27
\n
28x28
\n
\n \n \n
\n
0
\n
5
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
...
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
\n
\n
1
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
...
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
\n
\n
2
\n
4
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
...
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
\n
\n
3
\n
1
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
...
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
\n
\n
4
\n
9
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
...
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
0
\n
\n \n
\n
5 rows × 785 columns
\n
"
},
"metadata": {},
"execution_count": 34
}
],
"source": [
"mnist.head()"
]
},
{
"cell_type": "code",
"execution_count": 35,
"metadata": {},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"(60000, 784)"
]
},
"metadata": {},
"execution_count": 35
}
],
"source": [
"X_train, y_train = mnist.drop(['label'],axis=1).values, mnist[\"label\"]\n",
"X_train.shape"
]
},
{
"cell_type": "code",
"execution_count": 36,
"metadata": {},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"(60000,)"
]
},
"metadata": {},
"execution_count": 36
}
],
"source": [
"y_train.shape"
]
},
{
"cell_type": "code",
"execution_count": 37,
"metadata": {},
"outputs": [
{
"output_type": "display_data",
"data": {
"text/plain": "",
"image/svg+xml": "\n\n\n\n",
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAOcAAADnCAYAAADl9EEgAAAABHNCSVQICAgIfAhkiAAAAAlwSFlzAAALEgAACxIB0t1+/AAAADh0RVh0U29mdHdhcmUAbWF0cGxvdGxpYiB2ZXJzaW9uMy4xLjMsIGh0dHA6Ly9tYXRwbG90bGliLm9yZy+AADFEAAAFMUlEQVR4nO3dMUuVbRzH8c6jElGtIQgt7dEg1OA7CJwEF2fFNV1yipamcBEJwUDaegMOzrUkoi5iSKsZEg02SNB55gfO/T8Pp6P+Tn4+oz9uuZcvF3hxjq12u30DyPPPVb8A0Jk4IZQ4IZQ4IZQ4IdRwl92fcuHitTr90MkJocQJocQJocQJocQJocQJocQJocQJocQJocQJocQJocQJocQJocQJocQJocQJocQJocQJocQJocQJocQJocQJocQJocQJocQJocQJocQJocQJocQJocQJocQJocQJocQJocQJocQJocQJocQJocQJoYav+gUG0eHhYblvbGyU+/7+fuM2PT1dPnvv3r1y72ZiYqLcb9++/Ue/n/5xckIocUIocUIocUIocUIocUIocUIo95w9uHnzZrlvbm6W+8nJSePW7R7y2bNn5X56elrud+7cKfetra3G7cmTJ+Wz9JeTE0KJE0KJE0KJE0KJE0KJE0K12u12tZcjl+/NmzflvrS0VO4/fvwo98nJycbt3bt35bN3794tdxq1Ov3QyQmhxAmhxAmhxAmhxAmhxAmhxAmh3HP+ZT59+lTujx8/7vl3v3//vtynpqZ6/t3XnHtOGCTihFDihFDihFDihFDihFDihFC+GvMv8/Dhw3IfHx8v96Ghocat29d20l9OTgglTgglTgglTgglTgglTgglTgjlnvMvs76+Xu7b29vl/vbt28ZtdHS0p3eiN05OCCVOCCVOCCVOCCVOCCVOCCVOCOWec8AcHByU+8uXLy/pTbhoTk4IJU4IJU4IJU4IJU4IJU4I5SplwHS7Kjk5OSn3Vqvjf5sjkJMTQokTQokTQokTQokTQokTQokTQrXa7Xa1lyP9Nz8/X+5ra2vl/vv373Lv9vWWX758adxu3bpVPkvPOl4+OzkhlDghlDghlDghlDghlDghlDghlM9zhrl//365j4yMlPv5+Xm5f/v2rdwXFhYat9XV1fJZ+svJCaHECaHECaHECaHECaHECaHECaF8nvOa+fr1a7lX96yzs7PlsysrKz29Ez7PCQNFnBBKnBBKnBBKnBBKnBBKnBDK5zmvmW7fWzszM9O4ffz4sd+vQ8HJCaHECaHECaHECaHECaHECaFcpfAfT58+bdx2d3cv8U1wckIocUIocUIocUIocUIocUIocUIo95zXzK9fv8r96OiocVtaWur361BwckIocUIocUIocUIocUIocUIocUIo95zXzOnpabk/f/68cXv9+nW/X4eCkxNCiRNCiRNCiRNCiRNCiRNCiRNCtdrtdrWXI4Pn+Pi43MfGxhq3R48elc/u7Oz09E7caHX6oZMTQokTQokTQokTQokTQokTQvnI2DWzvr5+1a/A/+TkhFDihFDihFDihFDihFDihFDihFDuOa/A58+fG7fp6eny2Q8fPpT73t5euf/J11t2+8gY/eXkhFDihFDihFDihFDihFDihFDihFDuOa/A9+/fG7ejo6Py2QcPHpT72dlZuf/8+bPcl5eXG7e5ubnyWfrLyQmhxAmhxAmhxAmhxAmhxAmhxAmh/AvAMK9evSr3ra2tcu/2ec8XL16U++LiYuM2MjJSPkvP/AtAGCTihFDihFDihFDihFDihFDihFDuOeHqueeEQSJOCCVOCCVOCCVOCCVOCCVOCCVOCCVOCCVOCCVOCCVOCCVOCCVOCCVOCCVOCCVOCCVOCCVOCCVOCCVOCCVOCDXcZe/4lX3AxXNyQihxQihxQihxQihxQihxQqh/AR6SrNtAYCILAAAAAElFTkSuQmCC\n"
},
"metadata": {
"needs_background": "light"
}
}
],
"source": [
"some_digit = X_train[22000]\n",
"some_digit_image = some_digit.reshape(28, 28)\n",
"plt.imshow(some_digit_image, cmap = matplotlib.cm.binary,\n",
"interpolation=\"nearest\")\n",
"plt.axis(\"off\")\n",
"plt.show()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"##### Entwickeln Sie einen binären Klassifikator:\n",
"- Erstellen Sie den Zielvektor. Verden Sie hierfür den Befehl `y_train_my_number = (y == my_number)`. Die Aussage ist True bei allen \"my_number\" und False bei allen anderen Ziffern. "
]
},
{
"cell_type": "code",
"execution_count": 38,
"metadata": {},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"count 60000\n",
"unique 2\n",
"top False\n",
"freq 53735\n",
"Name: label, dtype: object"
]
},
"metadata": {},
"execution_count": 38
}
],
"source": [
"# Ihr Code ...\n",
"my_number = 7\n",
"y_train_my_number = (y_train == my_number)\n",
"y_train_my_number.describe()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"- Wählen Sie einen Klassifikator aus, trainieren diesen und geben Sie die Accuracy an. In diesem Fall ist es ratsam das *stochastische Gradientenverfahren* (SGD) als Klassifikator zu verwenden. Er ist in der Klasse SGDClassifier in Scikit-Learn enthalten. Dieser Klassifikator hat den Vorteil, sehr große Datensätze effizient zu bearbeiten."
]
},
{
"cell_type": "code",
"execution_count": 39,
"metadata": {},
"outputs": [
{
"output_type": "execute_result",
"data": {
"text/plain": [
"SGDClassifier(alpha=0.0001, average=False, class_weight=None,\n",
" early_stopping=False, epsilon=0.1, eta0=0.0, fit_intercept=True,\n",
" l1_ratio=0.15, learning_rate='optimal', loss='hinge',\n",
" max_iter=1000, n_iter_no_change=5, n_jobs=None, penalty='l2',\n",
" power_t=0.5, random_state=None, shuffle=True, tol=0.001,\n",
" validation_fraction=0.1, verbose=0, warm_start=False)"
]
},
"metadata": {},
"execution_count": 39
}
],
"source": [
"from sklearn.linear_model import SGDClassifier\n",
"\n",
"sgd_clf = SGDClassifier()\n",
"sgd_clf.fit(X_train,y_train_my_number)"
]
},
{
"cell_type": "code",
"execution_count": 40,
"metadata": {},
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"0.9826333333333334\n"
]
}
],
"source": [
"sgd_clf_predictions = sgd_clf.predict(X_train)\n",
"print(accuracy_score(y_train_my_number,sgd_clf_predictions))"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"- Vergleich Sie Ihr Ergebnis mit einen sehr primitiven Klassifikator, der einfach jedes Bild der Kategorie ``never_my_number`` zuordnet:"
]
},
{
"cell_type": "code",
"execution_count": 41,
"metadata": {},
"outputs": [],
"source": [
"from sklearn.base import BaseEstimator\n",
"class NeverMyNumberClassifier(BaseEstimator):\n",
" def fit(self, X, y=None):\n",
" pass\n",
" def predict(self, X):\n",
" return np.zeros((len(X), 1), dtype=bool)"
]
},
{
"cell_type": "code",
"execution_count": 42,
"metadata": {},
"outputs": [
{
"output_type": "stream",
"name": "stdout",
"text": [
"0.8955833333333333\n"
]
}
],
"source": [
"never_my_number_clf = NeverMyNumberClassifier()\n",
"never_my_number_clf.fit(X_train, y_train_my_number)\n",
"never_my_number_clf_predictions = never_my_number_clf.predict(X_train)\n",
"print(accuracy_score(y_train_my_number, never_my_number_clf_predictions))"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Versuchen Sie mit eigenen Worten zu erklären, warum diese Genauigkeit mit einem solchen primitiven Klassifikator erreicht wird. \n",
"\n",
"**Antwort:**\n",
"Die Genauigkeit kommt daher, dass die meisten Elemente in die Klasse \"not_my_number\" fallen. Daher ist die Vorhersage, dass die aktuell vorliegende Nummer nicht die gewünschte ist in den meisten Fällen richtig."
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"- Zur Bewertung der Güte eines Klassifikator existieren noch andere Metriken, die ihre eigenen Vor- und Nachteile haben. Welches sind die bekanntesten Metriken und warum könnte es wichtig sein, die richtige Metrik für die richtige Situation zu wählen?\n",
"\n",
"**Antwort:** Mithilfe der Anzahl der falschen bzw. richtigen positiven und negativen Vorhersagen lassen sich mehrere Aussagen über die Verhältnisse zwischen den Vorhersagen und der Realität machen. Dies würde z.B. für den Fall des \"Never MyNumberClassifier\" bedeuten, dass er alle true_positives $t_p$ falsch vorhersagt ($t_p=f_n$)"
]
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
"version": "3.7.6-final"
}
},
"nbformat": 4,
"nbformat_minor": 4
}