{
"cells": [
{
"cell_type": "markdown",
"id": "4e35c934",
"metadata": {},
"source": [
"# Predicting Heart Disease\n",
"\n",
"The World Health Organization (WHO) estimates that 17.9 million people die every year because of cardiovascular diseases (CVDs).\n",
"\n",
"There are multiple risk factors that could contribute to CVD in an individual such as unhealthy diet, lack of physical activity or mental illnesses. Being able to identify these risk factors in individuals early on could help prevent a lot of premature deaths.\n",
"\n",
"In this project, we will use the [Kaggle dataset](https://www.kaggle.com/datasets/fedesoriano/heart-failure-prediction) and build a K-Nearest Neighbors classifier to accurately predict the likelihood of a patient having a heart disease in the future. "
]
},
{
"cell_type": "code",
"execution_count": 1,
"id": "7a6956c6",
"metadata": {},
"outputs": [],
"source": [
"# import libraries\n",
"import pandas as pd\n",
"\n",
"from sklearn.preprocessing import MinMaxScaler\n",
"from sklearn.model_selection import train_test_split, GridSearchCV\n",
"from sklearn.metrics import accuracy_score\n",
"from sklearn.neighbors import KNeighborsClassifier\n",
"\n",
"import seaborn as sns\n",
"import matplotlib.pyplot as plt\n",
"%matplotlib inline"
]
},
{
"cell_type": "markdown",
"id": "6089c814",
"metadata": {},
"source": [
"## EDA: Descriptive Statistics\n",
"\n",
"We will start with exploring our dataset. As per the source, each patient has the following information collected about them:\n",
"\n",
"\n",
"1. `Age`: age of the patient [years]\n",
"2. `Sex`: sex of the patient [M: Male, F: Female]\n",
"3. `ChestPainType`: chest pain type [TA: Typical Angina, ATA: Atypical Angina, NAP: Non-Anginal Pain, ASY: Asymptomatic]\n",
"4. `RestingBP`: resting blood pressure [mm Hg]\n",
"5. `Cholesterol`: serum cholesterol [mm/dl]\n",
"6. `FastingBS`: fasting blood sugar [1: if FastingBS > 120 mg/dl, 0: otherwise]\n",
"7. `RestingECG`: resting electrocardiogram results [Normal: Normal, ST: having ST-T wave abnormality (T wave inversions and/or ST elevation or depression of > 0.05 mV), LVH: showing probable or definite left ventricular hypertrophy by Estes' criteria]\n",
"8. `MaxHR`: maximum heart rate achieved [Numeric value between 60 and 202]\n",
"9. `ExerciseAngina`: exercise-induced angina [Y: Yes, N: No]\n",
"10. `Oldpeak`: oldpeak = ST [Numeric value measured in depression]\n",
"11. `ST_Slope`: the slope of the peak exercise ST segment [Up: upsloping, Flat: flat, Down: downsloping]\n",
"12. `HeartDisease`: output class [1: heart disease, 0: Normal]"
]
},
{
"cell_type": "code",
"execution_count": 2,
"id": "d12aa5ff",
"metadata": {},
"outputs": [],
"source": [
"#load dataset\n",
"df = pd.read_csv(\"heart_disease_prediction.csv\")"
]
},
{
"cell_type": "code",
"execution_count": 3,
"id": "bc74cce6",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"
\n",
"\n",
"
\n",
" \n",
"
\n",
"
\n",
"
Age
\n",
"
Sex
\n",
"
ChestPainType
\n",
"
RestingBP
\n",
"
Cholesterol
\n",
"
FastingBS
\n",
"
RestingECG
\n",
"
MaxHR
\n",
"
ExerciseAngina
\n",
"
Oldpeak
\n",
"
ST_Slope
\n",
"
HeartDisease
\n",
"
\n",
" \n",
" \n",
"
\n",
"
0
\n",
"
40
\n",
"
M
\n",
"
ATA
\n",
"
140
\n",
"
289
\n",
"
0
\n",
"
Normal
\n",
"
172
\n",
"
N
\n",
"
0.0
\n",
"
Up
\n",
"
0
\n",
"
\n",
"
\n",
"
1
\n",
"
49
\n",
"
F
\n",
"
NAP
\n",
"
160
\n",
"
180
\n",
"
0
\n",
"
Normal
\n",
"
156
\n",
"
N
\n",
"
1.0
\n",
"
Flat
\n",
"
1
\n",
"
\n",
"
\n",
"
2
\n",
"
37
\n",
"
M
\n",
"
ATA
\n",
"
130
\n",
"
283
\n",
"
0
\n",
"
ST
\n",
"
98
\n",
"
N
\n",
"
0.0
\n",
"
Up
\n",
"
0
\n",
"
\n",
"
\n",
"
3
\n",
"
48
\n",
"
F
\n",
"
ASY
\n",
"
138
\n",
"
214
\n",
"
0
\n",
"
Normal
\n",
"
108
\n",
"
Y
\n",
"
1.5
\n",
"
Flat
\n",
"
1
\n",
"
\n",
"
\n",
"
4
\n",
"
54
\n",
"
M
\n",
"
NAP
\n",
"
150
\n",
"
195
\n",
"
0
\n",
"
Normal
\n",
"
122
\n",
"
N
\n",
"
0.0
\n",
"
Up
\n",
"
0
\n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
" Age Sex ChestPainType RestingBP Cholesterol FastingBS RestingECG MaxHR \\\n",
"0 40 M ATA 140 289 0 Normal 172 \n",
"1 49 F NAP 160 180 0 Normal 156 \n",
"2 37 M ATA 130 283 0 ST 98 \n",
"3 48 F ASY 138 214 0 Normal 108 \n",
"4 54 M NAP 150 195 0 Normal 122 \n",
"\n",
" ExerciseAngina Oldpeak ST_Slope HeartDisease \n",
"0 N 0.0 Up 0 \n",
"1 N 1.0 Flat 1 \n",
"2 N 0.0 Up 0 \n",
"3 Y 1.5 Flat 1 \n",
"4 N 0.0 Up 0 "
]
},
"execution_count": 3,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df.head()"
]
},
{
"cell_type": "markdown",
"id": "f9a59e51",
"metadata": {},
"source": [
"The dataset seems to contain both numerical and categorical features. Let's look at the datatype for each column."
]
},
{
"cell_type": "code",
"execution_count": 4,
"id": "344a302b",
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Age int64\n",
"Sex object\n",
"ChestPainType object\n",
"RestingBP int64\n",
"Cholesterol int64\n",
"FastingBS int64\n",
"RestingECG object\n",
"MaxHR int64\n",
"ExerciseAngina object\n",
"Oldpeak float64\n",
"ST_Slope object\n",
"HeartDisease int64\n",
"dtype: object\n"
]
},
{
"data": {
"text/plain": [
"int64 6\n",
"object 5\n",
"float64 1\n",
"dtype: int64"
]
},
"execution_count": 4,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"print(df.dtypes)\n",
"df.dtypes.value_counts()"
]
},
{
"cell_type": "markdown",
"id": "aab3ff7b",
"metadata": {},
"source": [
"`7` features in total are numerical while `5` are categorical. However, two of the numerical features, `FastingBS` and `HeartDisease` are categorical as well. \n",
"\n",
"We will focus on the numerical variables first."
]
},
{
"cell_type": "code",
"execution_count": 5,
"id": "1bd0d2c0",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"
\n",
"\n",
"
\n",
" \n",
"
\n",
"
\n",
"
Age
\n",
"
RestingBP
\n",
"
Cholesterol
\n",
"
FastingBS
\n",
"
MaxHR
\n",
"
Oldpeak
\n",
"
HeartDisease
\n",
"
\n",
" \n",
" \n",
"
\n",
"
count
\n",
"
918.000000
\n",
"
918.000000
\n",
"
918.000000
\n",
"
918.000000
\n",
"
918.000000
\n",
"
918.000000
\n",
"
918.000000
\n",
"
\n",
"
\n",
"
mean
\n",
"
53.510893
\n",
"
132.396514
\n",
"
198.799564
\n",
"
0.233115
\n",
"
136.809368
\n",
"
0.887364
\n",
"
0.553377
\n",
"
\n",
"
\n",
"
std
\n",
"
9.432617
\n",
"
18.514154
\n",
"
109.384145
\n",
"
0.423046
\n",
"
25.460334
\n",
"
1.066570
\n",
"
0.497414
\n",
"
\n",
"
\n",
"
min
\n",
"
28.000000
\n",
"
0.000000
\n",
"
0.000000
\n",
"
0.000000
\n",
"
60.000000
\n",
"
-2.600000
\n",
"
0.000000
\n",
"
\n",
"
\n",
"
25%
\n",
"
47.000000
\n",
"
120.000000
\n",
"
173.250000
\n",
"
0.000000
\n",
"
120.000000
\n",
"
0.000000
\n",
"
0.000000
\n",
"
\n",
"
\n",
"
50%
\n",
"
54.000000
\n",
"
130.000000
\n",
"
223.000000
\n",
"
0.000000
\n",
"
138.000000
\n",
"
0.600000
\n",
"
1.000000
\n",
"
\n",
"
\n",
"
75%
\n",
"
60.000000
\n",
"
140.000000
\n",
"
267.000000
\n",
"
0.000000
\n",
"
156.000000
\n",
"
1.500000
\n",
"
1.000000
\n",
"
\n",
"
\n",
"
max
\n",
"
77.000000
\n",
"
200.000000
\n",
"
603.000000
\n",
"
1.000000
\n",
"
202.000000
\n",
"
6.200000
\n",
"
1.000000
\n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
" Age RestingBP Cholesterol FastingBS MaxHR \\\n",
"count 918.000000 918.000000 918.000000 918.000000 918.000000 \n",
"mean 53.510893 132.396514 198.799564 0.233115 136.809368 \n",
"std 9.432617 18.514154 109.384145 0.423046 25.460334 \n",
"min 28.000000 0.000000 0.000000 0.000000 60.000000 \n",
"25% 47.000000 120.000000 173.250000 0.000000 120.000000 \n",
"50% 54.000000 130.000000 223.000000 0.000000 138.000000 \n",
"75% 60.000000 140.000000 267.000000 0.000000 156.000000 \n",
"max 77.000000 200.000000 603.000000 1.000000 202.000000 \n",
"\n",
" Oldpeak HeartDisease \n",
"count 918.000000 918.000000 \n",
"mean 0.887364 0.553377 \n",
"std 1.066570 0.497414 \n",
"min -2.600000 0.000000 \n",
"25% 0.000000 0.000000 \n",
"50% 0.600000 1.000000 \n",
"75% 1.500000 1.000000 \n",
"max 6.200000 1.000000 "
]
},
"execution_count": 5,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df.describe()"
]
},
{
"cell_type": "markdown",
"id": "cbc1f5d4",
"metadata": {},
"source": [
"From the table above, we can observe that:\n",
"\n",
"- The average age of patients is ~`53` years.\n",
"- The median for `Cholesterol` is higher than its mean by roughly `25` mm/dl, indicating that it could be a left-skewed distribution with a possibility of outliers skewing the distribution.\n",
"- `RestingBP` and `Cholesterol` have a minimum value of zero.\n",
"- There don't seem to be any missing values in these columns. But we will have to confirm it across the entire dataset as well.\n",
"\n",
"`RestingBP` can't be `0`. And, as per the [American Heart Association](https://www.heart.org/en/health-topics/cholesterol/about-cholesterol/what-your-cholesterol-levels-mean), serum cholesterol is a composite of different measurements. So, it is unlikely that `Cholesterol` would be `0` as well. We will have to clean both of these up later.\n",
"\n",
"Next, we will look at the categorical variables. It would also be beneficial to look at how the target feature, `HeartDisease`, is related to those categories. Before that, let's quickly check if there are any missing values in the dataset or not."
]
},
{
"cell_type": "code",
"execution_count": 6,
"id": "5c436572",
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"Age 0\n",
"Sex 0\n",
"ChestPainType 0\n",
"RestingBP 0\n",
"Cholesterol 0\n",
"FastingBS 0\n",
"RestingECG 0\n",
"MaxHR 0\n",
"ExerciseAngina 0\n",
"Oldpeak 0\n",
"ST_Slope 0\n",
"HeartDisease 0\n",
"dtype: int64"
]
},
"execution_count": 6,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df.isna().sum()"
]
},
{
"cell_type": "markdown",
"id": "847a31a8",
"metadata": {},
"source": [
"There are no missing values in this dataset!"
]
},
{
"cell_type": "markdown",
"id": "75988364",
"metadata": {},
"source": [
"## EDA: Categorical Data"
]
},
{
"cell_type": "markdown",
"id": "47977c45",
"metadata": {},
"source": [
"We identified that most of the categorical columns are all of dtype **object**."
]
},
{
"cell_type": "code",
"execution_count": 7,
"id": "304a72e7",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"
\n",
"\n",
"
\n",
" \n",
"
\n",
"
\n",
"
Sex
\n",
"
ChestPainType
\n",
"
RestingECG
\n",
"
ExerciseAngina
\n",
"
ST_Slope
\n",
"
\n",
" \n",
" \n",
"
\n",
"
count
\n",
"
918
\n",
"
918
\n",
"
918
\n",
"
918
\n",
"
918
\n",
"
\n",
"
\n",
"
unique
\n",
"
2
\n",
"
4
\n",
"
3
\n",
"
2
\n",
"
3
\n",
"
\n",
"
\n",
"
top
\n",
"
M
\n",
"
ASY
\n",
"
Normal
\n",
"
N
\n",
"
Flat
\n",
"
\n",
"
\n",
"
freq
\n",
"
725
\n",
"
496
\n",
"
552
\n",
"
547
\n",
"
460
\n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
" Sex ChestPainType RestingECG ExerciseAngina ST_Slope\n",
"count 918 918 918 918 918\n",
"unique 2 4 3 2 3\n",
"top M ASY Normal N Flat\n",
"freq 725 496 552 547 460"
]
},
"execution_count": 7,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df.describe(include=['object'])"
]
},
{
"cell_type": "markdown",
"id": "dfe32c99",
"metadata": {},
"source": [
"We can confirm that those columns are indeed categorical given the number of unique values in each of them. But, we can't gather much else. Also, `FastingBS` and `HeartDisease` are categorical as well since they only contain binary values. We can confirm that quickly as well."
]
},
{
"cell_type": "code",
"execution_count": 8,
"id": "99fc3fcd",
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
"(array([0, 1], dtype=int64), array([0, 1], dtype=int64))"
]
},
"execution_count": 8,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df[\"FastingBS\"].unique(), df[\"HeartDisease\"].unique() "
]
},
{
"cell_type": "markdown",
"id": "79253815",
"metadata": {},
"source": [
"Let's start looking at the categories in more detail."
]
},
{
"cell_type": "code",
"execution_count": 9,
"id": "5af0a32f",
"metadata": {},
"outputs": [
{
"data": {
"image/png": "iVBORw0KGgoAAAANSUhEUgAAA7MAAANcCAYAAACNDPg4AAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjUuMSwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy/YYfK9AAAACXBIWXMAAAsTAAALEwEAmpwYAACKDElEQVR4nOzdeZxddXn48c+TSTLZNzIJWUmEsIQtyIgiLiggqEgQBWNdYksbbXHB1oW0/tRq8yu/utUN29QtuEGqIpEqilGquIBBIpBAIBBIhoQkhCyEwCQz8/z+uCfhkkzCZLlz58583q/XvO45z/meM89lMnznud/v+Z7ITCRJkiRJqiW9qp2AJEmSJEn7y2JWkiRJklRzLGYlSZIkSTXHYlaSJEmSVHMsZiVJkiRJNad3tRM4GCNHjsxJkyZVOw1JUjdx++23P5aZDdXOo5bZN0uSDqV99c01XcxOmjSJRYsWVTsNSVI3EREPVzuHWmffLEk6lPbVNzvNWJIkSZJUcyxmJUmSJEk1x2JWkqQeJCIeioi7ImJxRCwqYiMi4qaIuL94HV7WfnZELI+IZRFxbvUylyTp2SxmJUnqeV6RmdMys7HYvwJYmJlTgIXFPhExFZgBHA+cB1wVEXXVSFiSpN1ZzEqSpOnAvGJ7HnBhWfyazGzOzBXAcuC0zk9PkqQ9WcxKktSzJPDziLg9ImYVsdGZuQageB1VxMcBq8rObSpizxIRsyJiUUQsWr9+fQVTlyTpGTX9aB5JkrTfzsjM1RExCrgpIu7dR9toJ5Z7BDLnAnMBGhsb9zguSVIlWMzu5tQPXl3tFCRu/9Tbq52CpG4qM1cXr+si4jpK04bXRsSYzFwTEWOAdUXzJmBC2enjgdWdmrDUw53xxTOqnUK38Nv3/LbaKagCnGYsSVIPEREDI2Lwzm3gVcDdwAJgZtFsJnB9sb0AmBER9RExGZgC3Na5WUuS1L6KFbMRcUyx7P/Ory0RcbnL/0uSVDWjgVsi4s+UitL/ycwbgSuBcyLifuCcYp/MXALMB5YCNwKXZWZrVTKXJGk3FZtmnJnLgGkAxTL+jwDX8czy/1dGxBXF/od3W/5/LPCLiDjaTlOSpEMjMx8ETm4nvgE4ay/nzAHmVDg1SZL2W2dNMz4LeCAzH8bl/yVJkiRJB6mzitkZwPeK7YNa/l+SJEmSpIoXsxHRF7gA+O/natpObI/l/X2WnSRJkiSpM0ZmXw38KTPXFvtri2X/OZDl/zNzbmY2ZmZjQ0NDBdOWJEmSJHVVnVHMvplnphiDy/9LkiRJkg5SxVYzBoiIAZSW+H9nWfhKYH5EXAqsBC6G0vL/EbFz+f8WXP5fkiRJkrQXFS1mM3MbcNhuMZf/lyRJkiQdlM5azViSJEmSpEPGYlaSJEmSVHMsZiVJkiRJNcdiVpIkSZJUcyxmJUmSJEk1x2JWkiRJklRzLGYlSZIkSTXHYlaSJEmSVHMsZiVJkiRJNcdiVpIkSZJUcyxmJUmSJEk1x2JWkiRJklRzLGYlSZIkSTXHYlaSJEmSVHMsZiVJkiRJNaeixWxEDIuI70fEvRFxT0ScHhEjIuKmiLi/eB1e1n52RCyPiGURcW4lc5MkSZIk1a5Kj8x+HrgxM48FTgbuAa4AFmbmFGBhsU9ETAVmAMcD5wFXRURdhfOTJEmSJNWgihWzETEEeBnwNYDM3J6Zm4DpwLyi2TzgwmJ7OnBNZjZn5gpgOXBapfKTJEmSJNWuSo7MPg9YD3wjIu6IiK9GxEBgdGauASheRxXtxwGrys5vKmLPEhGzImJRRCxav359BdOXJKl7ioi6om++odj3FiBJUs2pZDHbG3g+8JXMPAV4kmJK8V5EO7HcI5A5NzMbM7OxoaHh0GQqSVLP8j5Kt/7s5C1AkqSaU8litgloysxbi/3vUypu10bEGIDidV1Z+wll548HVlcwP0mSepyIGA+8FvhqWdhbgCRJNadixWxmPgqsiohjitBZwFJgATCziM0Eri+2FwAzIqI+IiYDU4DbKpWfJEk91L8DHwLaymIHdQuQJEnV0LvC138P8J2I6As8CPwlpQJ6fkRcCqwELgbIzCURMZ9SwdsCXJaZrRXOT5KkHiMizgfWZebtEXFmR05pJ7bHLUARMQuYBTBx4sSDSVGSpA6raDGbmYuBxnYOnbWX9nOAOZXMSZKkHuwM4IKIeA3QDxgSEd+muAUoM9ccyC1AmTkXmAvQ2Ni4R7ErSVIlVPo5s5IkqYvIzNmZOT4zJ1Fa2OmXmflWvAVIklSDKj3NWJIkdX1X4i1AkqQaYzErSVIPlJk3AzcX2xvwFiBJUo1xmrEkSZIkqeZYzEqSJEmSao7FrCRJkiSp5ljMSpIkSZJqjsWsJEmSJKnmWMxKkiRJkmqOxawkSZIkqeZYzEqSJEmSao7FrCRJkiSp5ljMSpIkSZJqjsWsJEmSJKnmVLSYjYiHIuKuiFgcEYuK2IiIuCki7i9eh5e1nx0RyyNiWUScW8ncJEmSJEm1qzNGZl+RmdMys7HYvwJYmJlTgIXFPhExFZgBHA+cB1wVEXWdkJ8kSZIkqcZUY5rxdGBesT0PuLAsfk1mNmfmCmA5cFrnpydJkiRJ6uoqXcwm8POIuD0iZhWx0Zm5BqB4HVXExwGrys5tKmLPEhGzImJRRCxav359BVOXJEmSJHVVvSt8/TMyc3VEjAJuioh799E22onlHoHMucBcgMbGxj2OS5IkSZK6v4qOzGbm6uJ1HXAdpWnDayNiDEDxuq5o3gRMKDt9PLC6kvlJkiRJkmpTxYrZiBgYEYN3bgOvAu4GFgAzi2YzgeuL7QXAjIioj4jJwBTgtkrlJ0mSJEmqXZWcZjwauC4idn6f72bmjRHxR2B+RFwKrAQuBsjMJRExH1gKtACXZWZrBfOTJEmSJNWoihWzmfkgcHI78Q3AWXs5Zw4wp1I5SZIkSZK6h2o8mkeSJEmSpINiMStJkiRJqjkWs5IkSZKkmmMxK0mSJEmqORazkiRJkqSaYzErSVIPERH9IuK2iPhzRCyJiH8u4iMi4qaIuL94HV52zuyIWB4RyyLi3OplL0nSs1nMSpLUczQDr8zMk4FpwHkR8SLgCmBhZk4BFhb7RMRUYAZwPHAecFVE1FUjcUmSdmcxK0lSD5ElW4vdPsVXAtOBeUV8HnBhsT0duCYzmzNzBbAcOK3zMpYkae8sZiVJ6kEioi4iFgPrgJsy81ZgdGauASheRxXNxwGryk5vKmKSJFWdxawkST1IZrZm5jRgPHBaRJywj+bR3iX2aBQxKyIWRcSi9evXH6JMJUnaN4tZSZJ6oMzcBNxM6V7YtRExBqB4XVc0awImlJ02HljdzrXmZmZjZjY2NDRUMm1JknbpUDEbEQs7EpMkSZ3jQPrmiGiIiGHFdn/gbOBeYAEws2g2E7i+2F4AzIiI+oiYDEwBbjskb0CSpIPUe18HI6IfMAAYWSzTv3O60RBgbIVzkyRJuznIvnkMMK9YkbgXMD8zb4iI3wPzI+JSYCVwMUBmLomI+cBSoAW4LDNbD/mbkiTpAOyzmAXeCVxOqXO8nWc6zC3AlyuXliRJ2osD7psz807glHbiG4Cz9nLOHGDOgacrSVJl7HOacWZ+PjMnAx/IzOdl5uTi6+TM/FJHvkGxauIdEXFDse+D2SVJOkCHom+WJKk7eK6RWQAy84sR8WJgUvk5mXl1B05/H3APpelP8MyD2a+MiCuK/Q/v9mD2scAvIuJopzNJkrSng+ybJUmqeR0qZiPiW8CRwGJgZ3GZwD47zIgYD7yW0vSkvy/C04Ezi+15lFZS/DBlD2YHVkTEzgez/75D70SSpB7kQPtmSZK6iw4Vs0AjMDUz93i23HP4d+BDwOCy2LMezB4R5Q9m/0NZOx/MLknS3h1o3yxJUrfQ0efM3g0cvj8XjojzgXWZeXtHT2kn5oPZJUlq3373zZIkdScdHZkdCSyNiNuA5p3BzLxgH+ecAVwQEa8B+gFDIuLbFA9mL0ZlD+jB7MBcgMbGRj+NliT1VAfSN0uS1G10tJj9+P5eODNnA7MBIuJMSqsuvjUiPkXpgexXsueD2b8bEZ+ltACUD2aXJGnvPl7tBCRJqqaOrmb8v4fwe16JD2aXJOmgHOK+WZKkmtPR1Yyf4Jn7V/sCfYAnM3PI3s96RmbeTGnVYh/MXuOefnwNK3581a795s3rGHvGRWx/YiObH1xM9Kqjftgojjjvr+ndbyDNm9ez9Buz6Td8DAADxx7JxHPeUaXsJan7ONi+WZL2R7Yld37qTvoO68tx7zyOJx95kgeufYC25jbqR9Qz5e1T6N2/VFrsPNb6dCsRwUkfOIlefTq6VI/UcR0dmS1fjZiIuJDSY3PUw/QbMYbjZn4SgGxr467/uJyhR51K88Y1jHvZxUSvOh7532tZe+sNjHv5mwCoHzpq1zmSpEPDvllSZ1pz8xr6H96f1qdLEyeXf285k6ZPYuiUoaz9/VpW/3I1E187kWxN7v/W/Ux52xQGjhvIjid3EHXtrfMqHbwD+ogkM38EvPLQpqJa88TKJdQPa6B+6EiGTDqR6FUHlEZft2/dWOXsJKlnsW+WVCnNG5vZuHQjo08fvSv29NqnGXJUaSLIsGOHsWHxBgA23buJAWMHMHDcQAD6DOxD9LKYVWV0dJrxRWW7vSg9286VhHu4jffeyvBjX7RH/LG7fsPwY58ZHNi+eT33XP1/qOvbn7EveQODxh/TmWlKUrdk3yyps6z44QqOuOAIWpufWc5mwJgBbLxrIyNOGsGGOzbQvKm0qPpT654iCJZetZQdW3cw8vkjGXf2uGqlrm6uo6sZv65suwV4CJh+yLNRzWhrbWHTA3cw9qUXPyu+5g8LiF69GHHciwHoM3AYJ7zzc/TuP4htj67ggeu/wNR3/F/q6vtXI21J6k7smyVV3ON3P06fwX0YNHEQm+/fvCt+5FuOZMX3V7DqxlWMOHEEvepKEz6zLdny4JbSfbJ9e7HkS0sYOGEgw44ZVqV3oO6so/fM/mWlE1Ft2bLiTgaMOoI+A4fuim24+xa2PLCYKZd8mIjSdJJevfvQq3cfAAYcPpn6oaN4euOjDDx8clXylqTuwr5ZUmd44sEn2HjXRm5fejttO9pofbqV+66+j6PffjTHX3Y8UBqN3bikdItZ/bB6hhw1hD6DSn//DZ86nCebnrSYVUV06J7ZiBgfEddFxLqIWBsRP4iI8ZVOTl3Xxnv+wIiyKcabV9zJ2tv+h+e9/nJ69anfFd+xbQvZ1gZA86Z1NG96lPqhDZ2eryR1N/bNkjrDERccQeMnGzn146dy9DuOZujRQzn67Uez/YntQGkktulnTYw+o3Q/7bDjhrFt9TZat7eSrcmW5Vvof7gz8lQZHZ1m/A3guxTPhAXeWsTOqURS6tradjSz5eG7mfiqd+yKNS38Fm2tLSz/708BzzyCZ2vTMtb89oelxaGiFxPOeQe9+w+qUuaS1K3YN0uqmsduf4xHf/MoAIedfBijXjQKgN4DejP2FWO589N3QpRGZkccP6Kaqaob62gx25CZ3yjb/2ZEXF6BfFQDevWp5+R3X/Ws2PF//al22w4/+gUMP/oFnZGWJPU09s2SOtXQKUMZOqV0i9nYM8cy9syx7bZreEEDDS9wJp4qr6OP5nksIt4aEXXF11uBDZVMTJIk7ZN9sySpR+toMftXwCXAo8Aa4I2AC09IklQ99s2SpB6to9OMPwnMzMyNABExAvg0pY5UkiR1PvtmSVKP1tGR2ZN2dpYAmfk4cEplUpIkSR1g3yxJ6tE6Wsz2iojhO3eKT387OqorSZIOPftmSVKP1tFO7zPA7yLi+0BSukdnTsWykiRJz8W+WZLUo3WomM3MqyNiEfBKIICLMnNpRTOTJEl7Zd8sSerpOjwdqeggO9xJRkQ/4NdAffF9vp+ZHyumQV0LTAIeAi4pW7xiNnAp0Aq8NzN/1tHvJ0lST7O/fbMkSd1JR++ZPRDNwCsz82RgGnBeRLwIuAJYmJlTgIXFPhExFZgBHA+cB1wVEXUVzE+SpB4lIiZExK8i4p6IWBIR7yviIyLipoi4v3gtvxd3dkQsj4hlEXFu9bKXJOnZKlbMZsnWYrdP8ZXAdGBeEZ8HXFhsTweuyczmzFwBLAdOq1R+kiT1QC3AP2TmccCLgMuKD5P9oFmSVHMqOTJLRNRFxGJgHXBTZt4KjM7MNQDF66ii+ThgVdnpTUVs92vOiohFEbFo/fr1lUxfkqRuJTPXZOafiu0ngHso9bV+0CxJqjkVXcI/M1uBaRExDLguIk7YR/No7xLtXHMuMBegsbFxj+OSJOm5RcQkSs+l3eOD5ogo/6D5D2WntftBs7qPlZ84sdop1LyJH72r2ilIPUZFR2Z3ysxNwM2UpiitjYgxAMXruqJZEzCh7LTxwOrOyE+SpJ4kIgYBPwAuz8wt+2raTmyPD5KdNSVJqoaKFbMR0VCMyBIR/YGzgXuBBcDMotlM4PpiewEwIyLqI2IyMAW4rVL5SZLUE0VEH0qF7Hcy84dF+KA+aM7MuZnZmJmNDQ0NlUtekqQylRyZHQP8KiLuBP5I6Z7ZG4ArgXMi4n7gnGKfzFwCzKf0iIEbgcuKacqSJOkQiIgAvgbck5mfLTvkB82SpJpTsXtmM/NOSvfi7B7fAJy1l3PmAHMqlZMkST3cGcDbgLuKBRoB/pHSB8vzI+JSYCVwMZQ+aI6InR80t+AHzZKkLqSiC0BJkqSuIzNvof37YMEPmiVJNaZTFoCSJEmSJOlQspiVJEmSJNUci1lJkiRJUs2xmJUkSZIk1RwXgJIkSZ3u1A9eXe0UuoXbP/X2aqcgSVXjyKwkSZIkqeZYzEqSJEmSao7FrCRJkiSp5ljMSpIkSZJqjsWsJEmSJKnmWMxKkiRJkmqOxawkSZIkqeZYzEqSJEmSak7FitmImBARv4qIeyJiSUS8r4iPiIibIuL+4nV42TmzI2J5RCyLiHMrlZskSZIkqbZVcmS2BfiHzDwOeBFwWURMBa4AFmbmFGBhsU9xbAZwPHAecFVE1FUwP0mSJElSjapYMZuZazLzT8X2E8A9wDhgOjCvaDYPuLDYng5ck5nNmbkCWA6cVqn8JEmSJEm1q1PumY2IScApwK3A6MxcA6WCFxhVNBsHrCo7ramI7X6tWRGxKCIWrV+/vqJ5S5IkSZK6pooXsxExCPgBcHlmbtlX03ZiuUcgc25mNmZmY0NDw6FKU5IkSZJUQ3pX8uIR0YdSIfudzPxhEV4bEWMyc01EjAHWFfEmYELZ6eOB1ZXMT9KBW/mJE6udgsTEj95V7RQkSVKVVHI14wC+BtyTmZ8tO7QAmFlszwSuL4vPiIj6iJgMTAFuq1R+kiRJkqTaVcmR2TOAtwF3RcTiIvaPwJXA/Ii4FFgJXAyQmUsiYj6wlNJKyJdlZmsF85MkSZIk1aiKFbOZeQvt3wcLcNZezpkDzKlUTpIkSZKk7qFTVjOWJEmSJOlQspiVJEmSJNWciq5mLEmSJHXU6s3bef8PH2H91hZ6BfzFqcP5q9NHsmlbC5f99yqaNu1g/LA+XHXJRIb2r2Nx0zZm/7j08ItMuPwVozjvuCFVfheSOovFrCRJPUREfB04H1iXmScUsRHAtcAk4CHgkszcWBybDVwKtALvzcyfVSFt9SB1vYKPnHs4J47tz9bmVs7/zwd4yZGD+P7iTZzxvEH83UsbuOo367nqN+uZ/arDOWZUP34860h61wVrn9jBq7+ynLOPHkzvur0t2yKpO3GasSRJPcc3gfN2i10BLMzMKcDCYp+ImArMAI4vzrkqIuo6L1X1RKMH9+HEsf0BGFRfx1Ej61n7RAs33buFN0wbBsAbpg3j5/duAaB/3167CtfmltzryqOSuidHZiVJ6iEy89cRMWm38HTgzGJ7HnAz8OEifk1mNgMrImI5cBrw+05JVj3eqo3bWfLo00wb15/Hnmxh9OA+QKngfezJll3t7mjaxgd/9AiPbN7B5y4a76is1IM4MitJUs82OjPXABSvo4r4OGBVWbumIiZV3JPNrbzr2pV89LzDGdxv3xMCThk/gF+8ewoLZj2Pq36znqd3tHVSlpKqzZFZSZLUnvaGt7LdhhGzgFkAEydOrGROh9TDN36VzQ8spveAIUz9y/8LwIM//jLNjz8KQGvzNurqB3DczE/y5JoHWPnzbxZnJmNefCHDpjRWJ/Fubkdr8q5rV3HhScN49dShAIwc2Ju1T+xg9OA+rH1iByMH7vkn7JSGfvTv04v71jVz0rj+nZ22pCqwmJUkqWdbGxFjMnNNRIwB1hXxJmBCWbvxwOr2LpCZc4G5AI2Nje0WvF3RiONfQsMpZ/PQT+buij3vdZft2m761feoqy8VRf1HjufYt32c6FXHjq2buGfeRxh65ClEL28jPpQykw9d/whHNdTzNy8euSt+9jFD+MHiTfzdSxv4weJNnHNsacXilRu3M3ZIH3rXBU2btvPghmbGD+tTrfQldTKLWUmSerYFwEzgyuL1+rL4dyPis8BYYApwW1UyrJDBE46lefP6do9lJhvvu40pl3wYgF596ncda2vZAeF9mZWwaOU2fvjnTRw7up5Xf2U5AB88azR/99KR/N38VVz7p42MHdqHr1wyoWj/JFf95jH61AUR8C+vHcuIdkZtJXVP/rZLktRDRMT3KC32NDIimoCPUSpi50fEpcBK4GKAzFwSEfOBpUALcFlmtlYl8SrY2rSMPgOG0G/44btiT655gIdv/Crbt2xg0mtmOSpbAS84YiAP//MJ7R773jsm7xG76OThXHTy8EqnJamLspiVJKmHyMw37+XQWXtpPweYU7mMuq6N9/6B4ce+6FmxgWOOZOpf/itPbVjNwz+dy5DJJ9Grd98qZShJcjVjSZKkMtnWyqb7b2f4sS9s93j/w8bSq089Tz32SCdnJkkqZzErSZJUZsvDS+g3Ygx9B4/YFWvetJ5sK82ybt78GM2PP0r9kJF7u4QkqRNUbJpxRHwdOB9Yl5knFLERwLXAJOAh4JLM3Fgcmw1cCrQC783Mn1UqN0mqtA/8qIlf3vcEhw3szU2XTQFg6aNP8Y8/Xs227W2MH9aXz79hPIP71bG4aRuzf1xaJDYTLn/FKM47bkg105d6hBU3XMUTq+6l5amt3PUflzPmjNcz8sSXs/HeW/eYYrz1kftYe90NRK/eEMGEs99O7wGDq5S5JAkqe8/sN4EvAVeXxa4AFmbmlRFxRbH/4YiYCswAjqe0YuIvIuLonrTQhKTu5eJpw5l52mH8/XVNu2Ifvn41/3Tu4bxo0kCu/dNG/vO3j/GBs0ZzzKh+/HjWkfSuC9Y+sYNXf2U5Zx89mN51rpYqVdLk8/+u3fikV//NHrHDjj+Dw44/o9IpSZL2Q8WmGWfmr4HHdwtPB+YV2/OAC8vi12Rmc2auAJYDp1UqN0mqtBdOGsiw/s9e6fTBDc288IgBALz0yIH89J4tAPTv22tX4drckljCSpIkPbfOvmd2dGauASheRxXxccCqsnZNRWwPETErIhZFxKL169t/NpwkdUVHj6rnpmVPAPA/S7awZvOOXcfuaNrG2V+6n3OvWs6c141zVFaSJOk5dJUFoNr7qy3ba5iZczOzMTMbGxoaKpyWJB06n5o+nqtv28Br/2M5Tza30aesYD1l/AB+8e4pLJj1PK76zXqe3tFWxUwlSZK6vs5+zuzaiBiTmWsiYgywrog3ARPK2o0HVndybpJUUUc11PPtt08G4MHHmvnl/U/s0WZKQz/69+nFfeuaOWlc/85OUZIkqWZ09sjsAmBmsT0TuL4sPiMi6iNiMjAFuK2Tc5OkinpsawsAbW3JF3+9nrc0lh77sXLjdlpaS5NRmjZt58ENzYwf1qdqeUqSJNWCSj6a53vAmcDIiGgCPgZcCcyPiEuBlcDFAJm5JCLmA0uBFuAyVzKWVMve89+r+P1DT7JxWwsv/My9vP/MUWzb3sbVfyyti3fecUO45JRhACxa+SRX/eYx+tQFEfAvrx3LiIGdPXFGkiSptlTsr6XMfPNeDp21l/ZzgDmVykeSOtMXL57QbvyvTh+5R+yik4dz0cnDK52SJElSt9JVFoCSJEmSJKnDLGYlSZIkSTXHm7IkSZIkaT/878teXu0Uat7Lf/2/B30NR2YlSZIkSTXHYlaSJEmSVHMsZiVJkiRJNcd7ZiVJkiSphrVm8s47/sTI+nquPP4EAH74yCNct2Y1dRG8aMQI3jX5eVXO8tCzmJUkSZKkGvaDRx7hiAEDeLK1FYA7Nm3ilsc38LXnn0rfXr3YuH17lTOsDKcZS5IkSVKNWtfczB8ef5zXHn74rtj1a1bzF+Mn0LdXqdwb3rdvtdKrKEdmJUmSJKlGfemBB3jn5MlsK0ZlAVY99RR3bdnM1x5+iL7Ri7993vM4dvDgKmZZGY7MSpIkSVIN+t2GDQzv24djditUWzN5oqWFq06exrueN5mP37OUzKxSlpXjyKwkSZIk1aC7t2zhtxs28IfHH2d7WxvbWlv5l3vvpaG+npceNpKI4LjBQ+gVweYdOxjWzaYbW8xKkiRJUg2aNXkysyZPBkqLPl37SBMfOfZYrl+zmjs2beKUYcNYtW0bO9raGNqnT5WzPfQsZiVJkiSpG3nN6MP5f/fdxztuX0Sf6MXsY44hIqqd1iHX5YrZiDgP+DxQB3w1M6+sckqSJPVo9s2S1PWdMmwYpwwbBkCfXr34yLHHVjehTtClFoCKiDrgy8CrganAmyNianWzkiSp57JvliR1VV2qmAVOA5Zn5oOZuR24Bphe5ZwkSerJ7JslSV1SVytmxwGryvabipgkSaoO+2ZJUpfU1e6Zbe+u5Gc9ECkiZgGzit2tEbGs4llpf40EHqt2ErUsPj2z2imoc/i7crA+dsgXszjiUF+wG+jJfXOX/x3tgf1Fl/+ZVOD/S7Wgy/9c4r097ufS5X8mdHxBqr32zV2tmG0CJpTtjwdWlzfIzLnA3M5MSvsnIhZlZmO185C6On9XVCN6bN/s72jX48+ka/Ln0vX0lJ9JV5tm/EdgSkRMjoi+wAxgQZVzkiSpJ7NvliR1SV1qZDYzWyLi3cDPKC3///XMXFLltCRJ6rHsmyVJXVWXKmYBMvMnwE+qnYcOSrebaiZViL8rqgk9uG/2d7Tr8WfSNflz6Xp6xM8kMvO5W0mSJEmS1IV0tXtmJUmSJEl6ThazOmgRkRHxrbL93hGxPiJuqGZeUlcVEa0Rsbjsa1K1c5J6qoh4fdGPHRsRtxa/kyuLfmzX72jRtz0WEf9a7Zy7q+Ln8Jmy/Q9ExMd3a/PniPjebrFvRsSK4mf1p4g4vZNS7jHKf0+K/V4R8YWIuDsi7oqIPxaLxH03Iv627LwXRsSdEdHlbm3sLiLisLL/Vz0aEY+U7Y+OiB0R8c5q51kpFrM6FJ4EToiI/sX+OcAjVcxH6uqeysxpZV8PVTshqQd7M3ALMCMzX5iZ04CPAtfu9jv6KmAZcElExx+OqP3SDFwUESPbOxgRx1H62/VlETFwt8MfLH52VwD/WdEse6ZdvyfF/puAscBJmXki8HpgE/B+4IMR0RARvYAvAX+XmS2dn3LPkJkbdv6/CvgP4HNl+28A/kDp59ctWczqUPkp8Npi+83A9/bRVpKkqouIQcAZwKU880f63rwZ+DywEnhRhVPrqVooLVrz/r0c/wvgW8DPgQv20ubXwFGHPrWeay+/J2OANZnZBpCZTZm5MTPXAp8G/g14F3BnZt5ShbRV8mbgH4DxETGu2slUgsWsDpVrgBkR0Q84Cbi1yvlIXVn/silA11U7GakHuxC4MTPvAx6PiOe316iYeXQWcAOlD2u77ShHF/Bl4C0RMbSdY28CrmXfP4PXAXdVKLee6kL2/D2ZD7yu6Mc+ExGnlLX/D2Aq8EHgQ52erQCIiAnA4Zl5G6Wf15uqnFJFWMzqkMjMO4FJlDqXnvj4Bml/lE8zfn21k5F6sDdT+jCW4nVvBdL5wK8ycxvwA+D1EVHXCfn1OJm5BbgaeG95PCJeAKzPzIeBhcDzI2J4WZNPRcRiYBalEUQdOnv8nmRmE3AMMBtoAxZGxFkAxWjtfwI/zcwNVchXJTMoFbGw7/+/1TRvxtahtIDS1JIzgcOqm4okSXsXEYcBr6S05kMCdUBGRHsjSW8GzoiIh4r9w4BXAL/ojFx7oH8H/gR8oyz2ZuDYsp/BEEr3A3612P9gZn6/sxLsKfb1e5KZzZRuM/tpRKylNIK7sDi1rfhS9bwZGB0Rbyn2x0bElMy8v5pJHWqOzOpQ+jrwicx0eo8kqat7I3B1Zh6RmZMycwKwAnhJeaOIGFLEJhbtJgGX0U1HObqCzHyc0ojSpVBaORe4mNJiQzt/BtPxZ9AZ9vZ78rKIGAu7fj4nAQ9XMU+ViYhjgIGZOa7sd+Zfee61AWqOxawOmeLm/89XOw9JkjrgzcDu96z/gNIiQ+UuAn5ZjELtdD1wQUTUVzC/nu4zwM5VjV8GPJKZ5U9K+DUwNSLGdHpmPcvefk++Cfw4Iu4G7qS0eNeXOjc17cPefm7d7gOgyMxq5yBJkiRJ0n5xZFaSJEmSVHMsZiVJkiRJNcdiVpIkSZJUcyxmJUmSJEk1x2JWkiRJklRzLGalbiIi/ikilkTEnRGxOCJeWO2cJEnqaiLi8Ii4JiIeiIilEfGTiJgVETccgmufGREvLtv/eEQ8UvTLd0fEBc9x/k8iYtg+jn+5uNbSiHiq2F4cEW882NylWtS72glIOngRcTpwPvD8zGyOiJFA3yqnJUlSlxIRQen5m/Myc0YRmwa87hB9izOBrcDvymKfy8xPR8RxwG8iYlRmtrV3cma+Zl8Xz8zLACJiEnBDZk47FElLtcqRWal7GAM8lpnNAJn5WGaujohTI+J/I+L2iPhZRIyJiKERsSwijgGIiO9FxN9UNXtJkjrHK4AdmfkfOwOZuRj4DTAoIr4fEfdGxHeKwpf2+tIi/t5ihPTOYqR3EvAu4P3FaOlLy79xZt4DtAAjI+JHxfWWRMSsnW0i4qGIGBkRkyLinoj4r6LNzyOif3tvKCK+FRHTy/a/ExEXRMQ7IuL6iLix6Pc/VtbmrRFxW5Hnf0ZE3UH/l5WqwGJW6h5+DkyIiPsi4qqIeHlE9AG+CLwxM08Fvg7MyczNwLuBb0bEDGB4Zv5X9VKXJKnTnADcvpdjpwCXA1OB5wFn7K0vLdpfAZySmScB78rMh4D/oDQSOy0zf1N+8eL2nzZgPfBXxfUagfdGxGHt5DMF+HJmHg9sAt6wl7y/Cvxl8T2GAi8GflIcOw14CzANuDgiGosR4jcBZxQju61FG6nmOM1Y6gYyc2tEnAq8lNKnztcC/0Kp076p+HC5DlhTtL8pIi4GvgycXJWkJUnqWm7LzCaAiFgMTKJURLbblwJ3At+JiB8BP9rHdd8fEW8FngDelJlZjOq+vjg+gVLhumG381YUo8ZQKsAntXfxzPzf4l7aUcBFwA8ys6XI96bM3FC8px8CL6E0Onwq8MeiTX9g3T7yl7osi1mpm8jMVuBm4OaIuAu4DFiSmafv3jYiegHHAU8BI4CmTkxVkqRqWQLsbbGk5rLtVkp/Jwd76UuB1wIvAy4A/k9EHL+X634uMz+9cycizgTOBk7PzG0RcTPQrwP5tDvNuPAtSqOrM4C/Kovnbu2S0nual5mz93E9qSY4zVjqBiLimIiYUhaaBtwDNBSLQxERfco62vcXx98MfL2YRiVJUnf3S6C+fK2IiHgB8PK9tF9GO31p8aHwhMz8FfAhYBgwiNLo6+DnyGEosLEoZI8FXnQwb6jwTUpTpMnMJWXxcyJiRHG/7YXAb4GFwBuLkVyK40ccghykTufIrNQ9DAK+WCzn3wIsB2YBc4EvFPfQ9Ab+PSJ2AH8NnJaZT0TEr4GPAB9r98qSJHUTxRTf11PqD68AngYeYi/ThDNze/HYm2f1pcB9wLeLWFAafd0UET8Gvl8syPSevaRxI/CuiLiTUrH8h0PwvtZGxD3tvI9bKI3aHgV8NzMXAUTER4CfF0X5DkqzuR4+2DykzhaZu88+kCRJklQrImIAcBelR/RtLmLvABoz893VzE2qJKcZS5IkSTUqIs4G7gW+uLOQlXoKR2YlSZIkSTXHkVlJkiRJUs2xmJUkSZIk1RyLWUmSJElSzanpR/OMHDkyJ02aVO00JEndxO233/5YZjZUO49aZt8sSTqU9tU313QxO2nSJBYtWlTtNCRJ3URE+JzFg2TfLEk6lPbVNzvNWJIkSZJUcyxmJUmSJEk1x2JWkiRJklRzLGYlSZIkSTWnYsVsRBwTEYvLvrZExOURMSIiboqI+4vX4WXnzI6I5RGxLCLOrVRukiRJkqTaVrFiNjOXZea0zJwGnApsA64DrgAWZuYUYGGxT0RMBWYAxwPnAVdFRF2l8pMkSZIk1a7OmmZ8FvBAZj4MTAfmFfF5wIXF9nTgmsxszswVwHLgtE7KT5IkSZJUQzrrObMzgO8V26Mzcw1AZq6JiFFFfBzwh7JzmopYpzr1g1d39reU9nD7p95e7RQkqVPZ/9YW+ylJXUHFR2Yjoi9wAfDfz9W0nVi2c71ZEbEoIhatX7/+UKQoSZIkSaoxnTHN+NXAnzJzbbG/NiLGABSv64p4EzCh7LzxwOrdL5aZczOzMTMbGxoaKpi2JEmSJKmr6oxi9s08M8UYYAEws9ieCVxfFp8REfURMRmYAtzWCflJkiRJkmpMRe+ZjYgBwDnAO8vCVwLzI+JSYCVwMUBmLomI+cBSoAW4LDNbK5mfJEmSJKk2VbSYzcxtwGG7xTZQWt24vfZzgDmVzEmSJEmSVPs669E8kiRJkiQdMhazkiRJkqSaYzErSZIkSao5FrOSJEmSpJpjMStJkiRJqjkWs5IkSZKkmmMxK0lSDxIRD0XEXRGxOCIWFbEREXFTRNxfvA4vaz87IpZHxLKIOLd6mUuS9GwWs5Ik9TyvyMxpmdlY7F8BLMzMKcDCYp+ImArMAI4HzgOuioi6aiQsSdLuLGYlSdJ0YF6xPQ+4sCx+TWY2Z+YKYDlwWuenJ0nSnixmJUnqWRL4eUTcHhGzitjozFwDULyOKuLjgFVl5zYVsWeJiFkRsSgiFq1fv76CqUuS9Ize1U5AkiR1qjMyc3VEjAJuioh799E22onlHoHMucBcgMbGxj2OS5JUCY7MSpLUg2Tm6uJ1HXAdpWnDayNiDEDxuq5o3gRMKDt9PLC687KVJGnvLGYlSeohImJgRAzeuQ28CrgbWADMLJrNBK4vthcAMyKiPiImA1OA2zo3a0mS2uc0Y0mSeo7RwHURAaW/Ab6bmTdGxB+B+RFxKbASuBggM5dExHxgKdACXJaZrdVJXZKkZ7OYlSSph8jMB4GT24lvAM7ayzlzgDkVTk2SpP1W0WnGETEsIr4fEfdGxD0RcboPZpckSZIkHaxK3zP7eeDGzDyW0ifB9+CD2SVJkiRJB6lixWxEDAFeBnwNIDO3Z+YmfDC7JEmSJOkgVXJk9nnAeuAbEXFHRHy1WDnRB7NLkiRJkg5KJYvZ3sDzga9k5inAkxRTiveiww9mz8zGzGxsaGg4NJlKkiRJkmpKJYvZJqApM28t9r9Pqbj1weySJEmSpINSsWI2Mx8FVkXEMUXoLErPqfPB7JIkSZKkg1Lp58y+B/hORPQFHgT+klIB7YPZJUmSJEkHrKLFbGYuBhrbOeSD2SVJkiRJB6zSz5mVJEmSJOmQs5iVJEmSJNUci1lJkiRJUs2xmJUkSZIk1RyLWUmSJElSzbGYlSRJkiTVHItZSZIkSVLNsZiVJEmSJNUci1lJkiRJUs2xmJUkSZIk1RyLWUmSJElSzbGYlSRJkiTVHItZSZIkSVLNsZiVJKmHiYi6iLgjIm4o9kdExE0RcX/xOrys7eyIWB4RyyLi3OplLUnSs1nMSpLU87wPuKds/wpgYWZOARYW+0TEVGAGcDxwHnBVRNR1cq6SJLWrosVsRDwUEXdFxOKIWFTE/PRXkqQqiYjxwGuBr5aFpwPziu15wIVl8WsyszkzVwDLgdM6KVVJkvapM0ZmX5GZ0zKzsdj3019Jkqrn34EPAW1lsdGZuQageB1VxMcBq8raNRWxZ4mIWRGxKCIWrV+/viJJS5K0u2pMM/bTX0mSqiAizgfWZebtHT2lnVjuEcicm5mNmdnY0NBwUDlKktRRlS5mE/h5RNweEbOKmJ/+SpJUHWcAF0TEQ8A1wCsj4tvA2ogYA1C8rivaNwETys4fD6zuvHQlSdq7ShezZ2Tm84FXA5dFxMv20dZPfyVJqqDMnJ2Z4zNzEqVbe36ZmW8FFgAzi2YzgeuL7QXAjIioj4jJwBTgtk5OW5KkdvWu5MUzc3Xxui4irqM0bXhtRIzJzDV++itJUpdwJTA/Ii4FVgIXA2TmkoiYDywFWoDLMrO1emlKkvSMio3MRsTAiBi8cxt4FXA3fvorSVLVZebNmXl+sb0hM8/KzCnF6+Nl7eZk5pGZeUxm/rR6GUuS9GyVHJkdDVwXETu/z3cz88aI+CN++itJkiRJOggVK2Yz80Hg5HbiG4Cz9nLOHGBOpXKSJEmSJHUP1Xg0jyRJkiRJB8ViVpIkSZJUcyxmJUmSJEk1x2JWkiRJklRzLGYlSZIkSTXHYlaSJEmSVHMsZiVJkiRJNcdiVpIkSZJUcyxmJUmSJEk1x2JWkiRJklRzLGYlSZIkSTXHYlaSJEmSVHMsZiVJkiRJNcdiVpIkSZJUcyxmJUmSJEk1p0PFbEQs7EhsL+fWRcQdEXFDsT8iIm6KiPuL1+FlbWdHxPKIWBYR53b0TUiS1NMcTN8sSVJ3sM9iNiL6RcQIYGREDC8K0RERMQkY28Hv8T7gnrL9K4CFmTkFWFjsExFTgRnA8cB5wFURUbdf70aSpG7uEPXNkiTVvOcamX0ncDtwbPG68+t64MvPdfGIGA+8FvhqWXg6MK/YngdcWBa/JjObM3MFsBw4rUPvQpKknuOg+mZJkrqL3vs6mJmfBz4fEe/JzC8ewPX/HfgQMLgsNjoz1xTXXxMRo4r4OOAPZe2aitizRMQsYBbAxIkTDyAlSZJq1yHomyVJ6hb2WczulJlfjIgXA5PKz8nMq/d2TkScD6zLzNsj4swOfJto71u3k8tcYC5AY2PjHsclSeoJDrBv7gf8Gqgvzvl+Zn6smLZ8bXGth4BLMnNjcc5s4FKgFXhvZv6sEu9HkqT91aFiNiK+BRwJLKbUmUGp0NxrhwmcAVwQEa8B+gFDIuLbwNqIGFOMyo4B1hXtm4AJZeePB1Z39I1IktSTHGDf3Ay8MjO3RkQf4JaI+ClwEaX1LK6MiCsorWfx4d3WsxgL/CIijs7M1r19A0mSOkuHilmgEZiamR0eCc3M2cBsgGJk9gOZ+daI+BQwE7iyeL2+OGUB8N2I+CylDnMKcFtHv58kST3MgfTNCWwtdvsUX0lp3Yozi/g84Gbgw5StZwGsiIid61n8/hDkL0nSQenoc2bvBg4/RN/zSuCciLgfOKfYJzOXAPOBpcCNwGV+8itJ0l4dUN9cPDJvMaWZUTdl5q3stp4FUL6exaqy09tdz0KSpGro6MjsSGBpRNxGaYoSAJl5QUdOzsybKX3KS2ZuAM7aS7s5wJwO5iRJUk92QH1z8UHxtIgYBlwXESfso3mH1rNwcUZJUjV0tJj9eCWTkCRJ++3jB3NyZm6KiJspPdv9oNazcHFGSVI1dHQ14/+tdCKSJKnjDqRvjogGYEdRyPYHzgb+H6V1K1zPQpJUUzq6mvETPDOtqC+lBSOezMwhlUpMXdPTj69hxY+v2rXfvHkdY8+4iBFTz2DFDVexffNj9B06ksmvu4ze/Qbuard9ywaWfmM2Y158IaNf8JpqpC5J3coB9s1jgHkRUUdp3Yz5mXlDRPwemB8RlwIrgYuhtJ5FROxcz6IF17PYw91z/4FeffsR0Yvo1Ytj3/bPrP7tdWy462Z69y/9KMa+9I0Mfd7JbHnobh75zXyytZWoq2P8y2cweOLU6r4BSaphHR2ZHVy+HxEXUlrNUD1MvxFjOG7mJwHItjbu+o/LGXrUqTx62/8weOJUDn/h+Tx66w2svfUGxr38TbvOa/rVdxky+aRqpS1J3c6B9M2ZeSdwSjtx17M4CEdfcgW9Bzzrx8GoU8/d48Pb3v0Hc+Tr30/fQcN5an0Ty3/wKU581+c7M1VJ6lY6uprxs2Tmj4BXHtpUVGueWLmE+mEN1A8dyeblf+Kw418CwGHHv4RNy/+0q92m+2+n79AG+h3mApiSVCn2zV3fgNFH0HfQcAD6jRxHW8sO2lp2VDkrSapdHZ1mfFHZbi9Kz7ZzgYcebuO9tzL82BcB0LJtC30GDQOgz6BhtGzbAkDr9mbW3vY/HHXxh1j7x59WK1VJ6nbsm7uO+7//KQhoOOkVjDz5FQCsv2MhG5b8loGHT2bcmW9+1q03AJvuW8SAUUfQq3efaqQsSd1CR1czfl3ZdgvwEKUHqauHamttYdMDdzD2pRfvs92a3/2QUaeeS13ffp2UmST1GPbNXcDRf/ER+g4azo4nt7D8+/9G/YgxNEx7JWNOnw4Bq2/5IY/c/D2OOO+vd53z1GNNPPLra5ly8QermLkk1b6O3jP7l5VORLVly4o7GTDqCPoMHApA7wFD2LF1E30GDWPH1k30HlBa9OLJNQ+y6b5FPPLr+bQ2b4MIoq4Po55/TjXTl6SaZ9/cNeycNtxn4BCGHnUq2x59kMETjt11fORJL+eBH35u1/72Jx7nweu/wKTXzKJ+2OhOz1eSupOOTjMeD3wROIPSFKZbgPdlZlMFc1MXtvGePzCimGIMMPTIU9iw5BYOf+H5bFhyC0OPej4Ax7z5n3a1Wf3b66jrW28hK0mHgH1z9bVubwbaqOvbn9btzTzx8N0cfvr0XR/uQmndiP4jxwPQ8vSTPPDDzzL2pRczaNzR1UtckrqJjk4z/gbwXYql+oG3FjGrkh6obUczWx6+m4mveseu2OEvPJ8VP/4yG+76NX2HHMbk111WvQQlqWewb66ylm2befD6LwCQba0MP+50hk4+iYd+8p9sW7cSgPqhI5l4TmkQff0dv6B541oe/f0CHv39AgCOeuMH6TPQJx1K0oHoaDHbkJnfKNv/ZkRcXoF8VAN69ann5Hdf9axY7/6DmHLJh/d53tgzXl/JtCSpp7FvrrL6YaM4bua/7BGf9Jp3ttt+zOnTS/fSSpIOiY4+muexiHhrRNQVX28FNlQyMUmStE/2zZKkHq2jxexfAZcAjwJrgDcCLjwhSVL12DdLknq0jk4z/iQwMzM3AkTECODTlDpSSZLU+eybJUk9WkdHZk/a2VkCZObjwCmVSUmSJHWAfbMkqUfraDHbKyKG79wpPv3d56huRPSLiNsi4s8RsSQi/nnnuRFxU0TcX7yWX3d2RCyPiGURce6BvCFJknqI/e6bJUnqTjra6X0G+F1EfJ/Ss+wuAeY8xznNwCszc2tE9AFuiYifAhcBCzPzyoi4ArgC+HBETAVmAMcDY4FfRMTRmdm6/29LkqRu70D6ZkmSuo0Ojcxm5tXAG4C1wHrgosz81nOck5m5tdjtU3wlMB2YV8TnARcW29OBazKzOTNXAMuB0zr+ViRJ6jkOpG+WJKk76fB0pMxcCizdn4tHRB1wO3AU8OXMvDUiRmfmmuKaayJiVNF8HPCHstObitju15wFzAKYOHHi/qQjSVK3ciB9syRJ3UVH75k9IJnZmpnTgPHAaRFxwj6aR3uXaOeaczOzMTMbGxoaDlGmkiRJkqRaUtFidqfM3ATcDJwHrI2IMQDF67qiWRMwoey08cDqzshPkiRJklRbKlbMRkRDRAwrtvsDZwP3AguAmUWzmcD1xfYCYEZE1EfEZGAKcFul8pMkSZIk1a5KLuE/BphX3DfbC5ifmTdExO+B+RFxKbASuBggM5dExHxK9/60AJe5krEkSZIkqT0VK2Yz807aeXh7Zm4AztrLOXPwsQKSJEmSpOfQKffMSpIkSZJ0KFnMSpIkSZJqjsWsJEk9RERMiIhfRcQ9EbEkIt5XxEdExE0RcX/xOrzsnNkRsTwilkXEudXLXpKkZ7OYlSSp52gB/iEzjwNeBFwWEVOBK4CFmTkFWFjsUxybARxP6fF6VxULO0qSVHUWs5Ik9RCZuSYz/1RsPwHcA4wDpgPzimbzgAuL7enANZnZnJkrgOXAaZ2atCRJe2ExK0lSDxQRkyg9deBWYHRmroFSwQuMKpqNA1aVndZUxHa/1qyIWBQRi9avX1/RvCVJ2sliVpKkHiYiBgE/AC7PzC37atpOLPcIZM7NzMbMbGxoaDhUaUqStE8Ws5Ik9SAR0YdSIfudzPxhEV4bEWOK42OAdUW8CZhQdvp4YHVn5SpJ0r5YzEqS1ENERABfA+7JzM+WHVoAzCy2ZwLXl8VnRER9REwGpgC3dVa+kiTtS+9qJyBJkjrNGcDbgLsiYnER+0fgSmB+RFwKrAQuBsjMJRExH1hKaSXkyzKztdOzlqR2nPHFM6qdgvbTb9/z20N6PYtZSZJ6iMy8hfbvgwU4ay/nzAHmVCwpSZIOkNOMJUmSJEk1x5FZSZIkdVkrP3FitVPQfpj40buqnYJ6EEdmJUmSJEk1x2JWkiRJklRzKjbNOCImAFcDhwNtwNzM/HxEjACuBSYBDwGXZObG4pzZwKVAK/DezPxZpfKTdHCc9qWuwOlskiT1XJUcmW0B/iEzjwNeBFwWEVOBK4CFmTkFWFjsUxybARwPnAdcFRF1FcxPkiRJklSjKlbMZuaazPxTsf0EcA8wDpgOzCuazQMuLLanA9dkZnNmrgCWA6dVKj9JkiRJUu3qlHtmI2IScApwKzA6M9dAqeAFRhXNxgGryk5rKmK7X2tWRCyKiEXr16+vaN6SJEmSpK6p4sVsRAwCfgBcnplb9tW0nVjuEcicm5mNmdnY0NBwqNKUJEmSJNWQihazEdGHUiH7ncz8YRFeGxFjiuNjgHVFvAmYUHb6eGB1JfOTJEmSJNWmihWzERHA14B7MvOzZYcWADOL7ZnA9WXxGRFRHxGTgSnAbZXKT5IkSZJUuyr2aB7gDOBtwF0RsbiI/SNwJTA/Ii4FVgIXA2TmkoiYDyyltBLyZZnZWsH8JEmSJEk1qmLFbGbeQvv3wQKctZdz5gBzKpWTJEmSJKl76JTVjCVJkiRJOpQsZiVJkiRJNcdiVpIkSZJUcyxmJUmSJEk1x2JWkiRJklRzLGYlSZIkSTXHYlaSJEmSVHMsZiVJkiRJNcdiVpIkSZJUc3pXOwFJktQ5IuLrwPnAusw8oYiNAK4FJgEPAZdk5sbi2GzgUqAVeG9m/qwKaUtSt7L8O8t5fMnj9Bnch1NmnwLAk01P8sC1D9DW0kb0Cp53yfMYfMRg2lrbeOB7D/DkqifJtqThBQ2Mf9X4Kr+DrsORWUmSeo5vAuftFrsCWJiZU4CFxT4RMRWYARxfnHNVRNR1XqqS1D01vLCBqX879Vmxh65/iAmvnsC0D09j4msm8vD1DwOw4Y4NtLW0MW32NE764Ems/d1ant7wdDXS7pIcmZUkqYfIzF9HxKTdwtOBM4vtecDNwIeL+DWZ2QysiIjlwGnA7zslWanGfeBHTfzyvic4bGBvbrpsCgCXzV/Jgxu2A7Dl6VaG9Kvjp397FDtakw9f/wh3r3mKljZ4w8nDuOxlDdVMXxU09KihexSkEUHr060AtDzdQt+hfYsD0NbcRrYmbTvaiLqgrp+fK+5kMStJUs82OjPXAGTmmogYVcTHAX8oa9dUxPYQEbOAWQATJ06sYKpS7bh42nBmnnYYf39d067Yly955vfjkzeuYUhRlPzPks1sb01+ftkUntrextlfvp8LThzKhOF9Oz1vVcekiyax9CtLeehHD0HCCe8/AYDDph3G43c9zh8/8kfadrQx6fWT6DOwT3WT7UKcZixJFbB683be9I0VvPKL93P2l+7n679/DCj9wXL2l+5n0sfv5s5HntrjvEc2bee4OUv5z98+1tkpS7uLdmLZXsPMnJuZjZnZ2NDgaJIE8MJJAxnWv/0RtMzkf5Zs5oIThwIQAdu2t9HSmjzd0kafumBwvX+m9ySP3vIok18/mcZPNDLp9ZN44LsPALD14a1EBI3/0sjzP/Z8Vv9qNU8/5jTjnfwtkaQKqOsVfOTcw/nle6bwo795Hlf/8XHuW/c0R4+q5z9nTOSFRwxo97xP3PgoZx41qJOzVQ+3NiLGABSv64p4EzChrN14YHUn5yZ1S7c9vI2Rg3oz+bB6AF4zdSgD+vbiBZ++l9M/u4xZLx7JsAFOoOxJ1t+2nhEnjwDgsFMOY+vDWwF4bNFjDDtuGL3qetF3cF+GTB7C1pVbq5lql1KxYjYivh4R6yLi7rLYiIi4KSLuL16Hlx2bHRHLI2JZRJxbqbwkqTOMHtyHE8f2B2BQfR1Hjaxn7RMtTGnox5Ej69s952f3bGHi8L4cPar941KFLABmFtszgevL4jMioj4iJgNTgNuqkJ/U7Sy4azMXnDBs1/7iR7bRK+C2DxzLLZcfw3/97jFWPr69egmq0/Ud2pcty7cAsPm+zfRr6FeKD+/L5vs3k5m0NrfyxENP0H90/2qm2qVUcmT2m7hioiSxauN2ljz6NNPG7b3z2ba9ja/csp7Lz3SKpionIr5HaQGnYyKiKSIuBa4EzomI+4Fzin0ycwkwH1gK3Ahclpmt1clc6j5aWpMb79nM604Yuit2/Z2bOXPKIPrUBSMH9ebUiQO4c/Wet6Koe7jvm/dx1+fu4um1T7Po/yxi7e/XcuSMI3nouodYfOViVt6wkiNnHAnAmJeNoa25jcX/upg7P30no140ioHjBlb5HXQdFZu/4IqJkgRPNrfyrmtX8tHzDmfwPlYf/Oyv1vLXp49kYL2f46lyMvPNezl01l7azwHmVC4jqee55cGtHDmynjFDn1nEZ9zQPvzuwSd5/UnDeGpHckfTU1z6opFVzFKVdPQ7jm43fvKHTt4jVldfxzF/dUylU6pZnT0Z3xUTJfUYO1qTd127igtPGsarpw7dZ9vFTU/x06Vb+NebHmXL061EBPW9g3e88LBOylaSdCi9579X8fuHnmTjthZe+Jl7ef+Zo5hx6gh+fPdmLjhx2LPavv20EXzgR49wzpeXk8DF04Zx3OH9qpK3VEu6yp3l+7ViIjAXoLGxsd02klRtmcmHrn+Eoxrq+ZsXP/en69+/9Hm7tj/3q7UM6FtnIStJNeyLF09oN/6Z14/fIzawvo6vvMlBGml/dXYxuzYixhSjsq6YKKnbWrRyGz/88yaOHV3Pq7+yHIAPnjWa7a3Jx36ymsefbOUvv/MQUw/vz7fePqm6yUqSJNWgzi5md66YeCV7rpj43Yj4LDAWV0yUVONecMRAHv7nE9o9dt5xQ/Z57vtfMboSKUmSJHUrFStmixUTzwRGRkQT8DFKRez8YvXElcDFUFoxMSJ2rpjYgismSpIkSZL2oZKrGbtioiRJkiSpIir5nFlJkiRJkirCYlaSJEmSVHMsZiVJkiRJNcdiVpIkSZJUcyxmJUmSJEk1x2JWkiRJklRzLGYlSZIkSTXHYlaSJEmSVHMsZiVJkiRJNcdiVpIkSZJUcyxmJUmSJEk1x2JWkiRJklRzLGYlSZIkSTXHYlaSJEmSVHMsZiVJkiRJNafLFbMRcV5ELIuI5RFxRbXzkSSpp7NvliR1RV2qmI2IOuDLwKuBqcCbI2JqdbOSJKnnsm+WJHVVXaqYBU4Dlmfmg5m5HbgGmF7lnCRJ6snsmyVJXVLvaiewm3HAqrL9JuCF5Q0iYhYwq9jdGhHLOik3ddxI4LFqJ1HL4tMzq52COoe/KwfrY3Gor3jEob5gN2DffHC65e+5/dQh0S3/bVTg/8s9Uff8twHEew/o38de++auVsy29+7yWTuZc4G5nZOODkRELMrMxmrnIXV1/q6oRtg3HwR/z7U3/tvQ3vhvo+O62jTjJmBC2f54YHWVcpEkSfbNkqQuqqsVs38EpkTE5IjoC8wAFlQ5J0mSejL7ZklSl9SlphlnZktEvBv4GVAHfD0zl1Q5Le0/p5pJHePviro8++aD5u+59sZ/G9ob/210UGTmc7eSJEmSJKkL6WrTjCVJkiRJek4Ws5IkSZKkmmMxq0MmIs6LiGURsTwirqh2PlJXFRFfj4h1EXF3tXORtHcRkRHxmbL9D0TExzs5h5sjwkd0dAMR8U8RsSQi7oyIxRHxq+J1eURsLrYXR8SLq52rKisitpZtvyMivrfb8ZERsT4i6nf/f0BETPLvh2dYzOqQiIg64MvAq4GpwJsjYmp1s5K6rG8C51U7CUnPqRm4KCJGHsjJEdGlFtpU9UTE6cD5wPMz8yTgbOAtmTkN+GvgN5k5rfj6XRVTVef7IXBORAwoi70RWJCZzVXKqWZYzOpQOQ1YnpkPZuZ24BpgepVzkrqkzPw18Hi185D0nFoorSr6/t0PRMQREbGwGGVbGBETi/g3I+KzEfEr4P8V+18pRuEejIiXF7Mz7omIb5Zd7ysRsagYufvnznqD6jRjgMd2FieZ+Vhm+rxmkZlbgF8DrysLzwC+1/4ZKmcxq0NlHLCqbL+piEmSVMu+DLwlIobuFv8ScHUxyvYd4Atlx44Gzs7Mfyj2hwOvpFQU/xj4HHA8cGJETCva/FNmNgInAS+PiJMq8WZUNT8HJkTEfRFxVUS8vNoJqUv5HqUClogYS+n/Ib8qO/6dndPQgZ90fnpdl8WsDpVoJ+ZznyRJNa0YNbkaeO9uh04Hvltsfwt4Sdmx/87M1rL9H2fpWYh3AWsz867MbAOWAJOKNpdExJ+AOygVut6q041k5lbgVGAWsB64NiLeUdWk1JXcALwkIoYAlwDf3+3/IW/ZOQ0deE01EuyqLGZ1qDQBE8r2xwNOn5EkdQf/DlwKDNxHm/IPcJ/c7djO+97ayrZ37veOiMnAB4CzipHe/wH6HUzC6noyszUzb87MjwHvBt5Q7ZzUNWTmU8CNwOtxivF+sZjVofJHYEpETI6IvpR+ERdUOSdJkg5aZj4OzKdU0O70O4ppgcBbgFsO4lsMoVQAb46I0ZQWU1Q3EhHHRMSUstA04OEqpaOu6XvA3wOjgT9UOZeaYTGrQyIzWyh9yvgz4B5gfmYuqW5WUtdULMH/e+CYiGiKiEuf6xxJVfcZoHxV4/cCfxkRdwJvA953oBfOzD9Tml68BPg68NuDyFNd0yBgXkQsLf7NTAU+Xt2UVEUDiv5/59ffU7qveixwbXFbgjog/G8lSZIkSao1jsxKkiRJkmqOxawkSZIkqeZYzEqSJEmSao7FrCRJkiSp5ljMSpIkSZJqjsWs1EVERGtELC77mrSf518eEQPK9n8SEcMOMJczI2JzkcedEfGLiBhVHDsmIm4ujt0TEXMP5HtIktSVlPXDd0fEjw+kD42IaRHxmrL9CyLiioPI6eaIWFb2t8H3y469vch1SfHInw+UHfv7iLg3Iu6KiD9HxGcjos+B5iF1VRazUtfxVGZOK/t6aD/PvxzYVcxm5msyc9NB5PObIo+TgD8ClxXxLwCfK44dB3zxIL6HJEldxc5++ATgcZ7p9/bHNGBXMZuZCzLzyoPM6y1lfxu8ESAiXk2p339VZh4PPB/YXBx7F/Aq4EWZeSLwAmAd0P8g85C6nN7VTkBS+yJiEHA9MBzoA3wkM6+PiIHAfGA8UAd8EhhN6UHbv4qIxzLzFRHxENBI6UHtPwVuAV4MPAJMz8ynIuIFwNeAJ4vjry468fI8AhgMLC9CY4Cmnccz864KvH1Jkqrp98BJABFxJPBloAHYBvxNZt4bERcDHwNaKRWSZwOfAPpHxEuAf6VUQDZm5rsj4pvAFkp98+HAhzLz+xHRC/gS8HJgBaXBpq9n5q5R2HbMBj6QmasBMvNp4L+KY/8EvGznB9qZuR042IJa6pIcmZW6jv5l04iuA54GXp+ZzwdeAXymKCzPA1Zn5slF4XljZn4BWA28IjNf0c61pwBfLj693QS8oYh/A3hXZp5OqTMu99KIWAyspNRBf72Ifw74ZUT8NCLef6BTmSVJ6ooiog44C1hQhOYC78nMU4EPAFcV8Y8C52bmycAFRdH4UeDaYhT12nYuPwZ4CXA+zxSYFwGTgBOBvwZO3+2c75T9ffCpInYCcHs7uQ8GBmXmiv1821JNspiVuo7yacavBwL4vxFxJ/ALYBylEdi7gLMj4v9FxEszc3MHrr0iMxcX27cDk4oidHBm/q6If3e3c3ZOM55Aqej9N4DM/AZwHPDfwJnAHyKi/sDesiRJXUb/4kPcDcAI4KZiltSLgf8ujv0npYIU4LfANyPibyjNlOqIH2VmW2YupdSnQ6m4/e8i/ijwq93OKZ9m/MHnuH4AuWsn4tyiCH4oIl7cwRylmmExK3Vdb6E0penUzJwGrAX6ZeZ9wKmUitp/jYiPduBazWXbrZRuMYj9yGUB8LKdO5m5OjO/npnTgRZKnxBLklTLnir62yOAvpTume0FbNptTYvjADLzXcBHgAnA4og4rAPfo7w/jt1e98cSSn8LPEtmbgGejIjJxf7Pivd0d/GepG7FYlbquoYC6zJzR0S8glLnSkSMBbZl5reBT1Na9AHgCUr3tnZIZm4EnoiIFxWhGfto/hLggeL7n7dzRcSIOBw4jNJ9uJIk1bxixtN7KU0pfgpYUdwfS5ScXGwfmZm3ZuZHgccoFbX71RcXbgHeEBG9ImI0pVlPz+VfgX8r+mEioj4i3lt27Cs7bwMqblHqt585STXBBaCkrus7wI8jYhGwGLi3iJ8IfCoi2oAdwN8W8bnATyNizV7um23PpcB/RcSTwM0UKyEWdt4zG0X8r4v4q4DPR8TTxf4Hi2lRkiR1C5l5R0T8mdIHvW+hVBx+hNKCjNcAf6bUF0+h1E8uLGIrgSuK/vNfO/jtfkDpHt27gfuAW3l2f/ydiHiq2H4sM8/OzJ8Uhe8vimI1eWZti69QerrBrRHRDGylNCX6jv38zyB1eZGZz91KUrcUEYMyc2uxfQUwJjPfV+W0JEnqUXb2x8VU5duAM/ygWHpujsxKPdtrI2I2pf8XPAy8o7rpSJLUI91QTAvuC3zSQlbqGEdmJUmSJEk1xwWgJEmSJEk1x2JWkiRJklRzLGYlSZIkSTWnpheAGjlyZE6aNKnaaUiSuonbb7/9scxsqHYetcy+WZJ0KO2rb67pYnbSpEksWrSo2mlIkrqJiHi42jnUOvtmSdKhtK++2WnGkiRJkqSaYzErSZIkSao5FrOSJEmSpJpjMStJkiRJqjkWs5IkSZKkmmMxK0mSJEmqOTX9aJ5KOPWDV1c7BYnbP/X2aqcgSVKXsPITJ1Y7Be2HiR+9q9opqAdxZFaSJEmSVHMsZiVJkiRJNcdiVpIkSZJUcyxmJUmSJEk1x2JWkiRJklRzLGYlSZIkSTXHYlaSJEmSVHMsZiVJkiRJNcdiVpIkSZJUcyxmJUmSJEk1x2JWkiRJklRzKlrMRsRDEXFXRCyOiEVFbERE3BQR9xevw8vaz46I5RGxLCLOrWRukiRJkqTa1Rkjs6/IzGmZ2VjsXwEszMwpwMJin4iYCswAjgfOA66KiLpOyE+SJEmSVGOqMc14OjCv2J4HXFgWvyYzmzNzBbAcOK3z05MkSZIkdXWVLmYT+HlE3B4Rs4rY6MxcA1C8jiri44BVZec2FbFniYhZEbEoIhatX7++gqlLkiRJkrqq3hW+/hmZuToiRgE3RcS9+2gb7cRyj0DmXGAuQGNj4x7HJUmSJEndX0VHZjNzdfG6DriO0rThtRExBqB4XVc0bwImlJ0+HlhdyfwkSZIkSbWpYsVsRAyMiME7t4FXAXcDC4CZRbOZwPXF9gJgRkTUR8RkYApwW6XykyRJkiTVrkqOzI4GbomIP1MqSv8nM28ErgTOiYj7gXOKfTJzCTAfWArcCFyWma0VzE+SpB4pIuoi4o6IuKHY97F5kqSaU7F7ZjPzQeDkduIbgLP2cs4cYE6lcpIkSQC8D7gHGFLs73xs3pURcUWx/+HdHps3FvhFRBzth82SpK6gGo/mkSRJVRIR44HXAl8tC/vYPElSzbGYlSSpZ/l34ENAW1nMx+ZJkmqOxawkST1ERJwPrMvM2zt6Sjuxdh+bl5mNmdnY0NBwUDlKktRRlX7OrCRJ6jrOAC6IiNcA/YAhEfFtisfmZeaaaj0279QPXl2Jy6pCbv/U26udgiQ5MitJUk+RmbMzc3xmTqK0sNMvM/Ot+Ng8SVINcmRWkiRdCcyPiEuBlcDFUHpsXkTsfGxeCz42T5LUhVjMSpLUA2XmzcDNxbaPzZMk1RynGUuSJEmSao7FrCRJkiSp5ljMSpIkSZJqjsWsJEmSJKnmWMxKkiRJkmqOxawkSZIkqeZYzEqSJEmSao7FrCRJkiSp5ljMSpIkSZJqjsWsJEmSJKnmWMxKkiRJkmqOxawkSZIkqeZYzEqSJEmSao7FrCRJkiSp5lS8mI2Iuoi4IyJuKPZHRMRNEXF/8Tq8rO3siFgeEcsi4txK5yZJkiRJqk2dMTL7PuCesv0rgIWZOQVYWOwTEVOBGcDxwHnAVRFR1wn5SZIkSZJqTEWL2YgYD7wW+GpZeDowr9ieB1xYFr8mM5szcwWwHDitkvlJkiRJkmpTpUdm/x34ENBWFhudmWsAitdRRXwcsKqsXVMRkyRJkiTpWSpWzEbE+cC6zLy9o6e0E8t2rjsrIhZFxKL169cfVI6SJEmSpNpUyZHZM4ALIuIh4BrglRHxbWBtRIwBKF7XFe2bgAll548HVu9+0cycm5mNmdnY0NBQwfQlSZIkSV1VxYrZzJydmeMzcxKlhZ1+mZlvBRYAM4tmM4Hri+0FwIyIqI+IycAU4LZK5SdJkiRJql29q/A9rwTmR8SlwErgYoDMXBIR84GlQAtwWWa2ViE/SZIkSVIX1ynFbGbeDNxcbG8AztpLuznAnM7ISZIkSZJUuzrjObOSJEmSJB1SFrOSJEmSpJpjMStJkiRJqjkWs5IkSZKkmlON1YwldQMrP3FitVOQmPjRu6qdgiRJqhJHZiVJ6iEiol9E3BYRf46IJRHxz0V8RETcFBH3F6/Dy86ZHRHLI2JZRJxbvewlSXo2i1lJknqOZuCVmXkyMA04LyJeBFwBLMzMKcDCYp+ImArMAI4HzgOuioi6aiQuSdLuLGYlSeohsmRrsdun+EpgOjCviM8DLiy2pwPXZGZzZq4AlgOndV7GkiTtncWsJEk9SETURcRiYB1wU2beCozOzDUAxeuoovk4YFXZ6U1FbPdrzoqIRRGxaP369RXNX5KknSxmJUnqQTKzNTOnAeOB0yLihH00j/Yu0c4152ZmY2Y2NjQ0HKJMJUnaN4tZSZJ6oMzcBNxM6V7YtRExBqB4XVc0awImlJ02HljdeVlKkrR3FrOSJPUQEdEQEcOK7f7A2cC9wAJgZtFsJnB9sb0AmBER9RExGZgC3NapSUuStBc+Z1aSpJ5jDDCvWJG4FzA/M2+IiN8D8yPiUmAlcDFAZi6JiPnAUqAFuCwzW6uUuyRJz2IxK0lSD5GZdwKntBPfAJy1l3PmAHMqnJokSfvNacaSJEmSpJpjMStJkiRJqjkWs5IkSZKkmmMxK0mSJEmqOR0qZiNiYUdikiSpc9g3S5J6un2uZhwR/YABwMiIGA5EcWgIMLbCuUmSpN3YN0uSVPJcj+Z5J3A5pc7xdp7pMLcAX65cWpIkaS/smyVJ4jmK2cz8PPD5iHhPZn6xk3KSJEl7Yd8sSVLJc43MApCZX4yIFwOTys/JzKv3dk4xDerXQH1xzvcz82MRMQK4trjWQ8AlmbmxOGc2cCnQCrw3M3+2/29JkqTu70D6ZkmSupMOFbMR8S3gSGAxpUITIIF9dZjNwCszc2tE9AFuiYifAhcBCzPzyoi4ArgC+HBETAVmAMdTmjr1i4g4OjNb9/YNJEnqqQ6wb5YkqdvoUDELNAJTMzM7euGi7dZit0/xlcB04MwiPg+4GfhwEb8mM5uBFRGxHDgN+H1Hv6ckST3IfvfNkiR1Jx19zuzdwOH7e/GIqIuIxcA64KbMvBUYnZlrAIrXUUXzccCqstObitju15wVEYsiYtH69ev3NyVJkrqLA+qbJUnqLjo6MjsSWBoRt1GaPgxAZl6wr5OKKcLTImIYcF1EnLCP5tFObI9PmzNzLjAXoLGx0U+jJUk91QH1zZIkdRcdLWY/fjDfJDM3RcTNwHnA2ogYk5lrImIMpVFbKI3ETig7bTyw+mC+ryRJ3djHq52AJEnV1NHVjP93fy8cEQ3AjqKQ7Q+cDfw/YAEwE7iyeL2+OGUB8N2I+CylBaCmALft7/eVJKknOJC+WZKk7qSjqxk/wTNTfvtSWszpycwcso/TxgDzIqKO0r258zPzhoj4PTA/Ii4FVgIXA2TmkoiYDywFWoDLXMm4a7p77j/Qq28/InoRvXpx7Nv+edextX/8CY/877Wc9HdfoveAwTy+9Hes/eNPdx1/av0qjn37PzNg1BHVSF2Suo0D7JsldaLWtuT8/3yAw4f04RtvKf3t840/bODq2zZQ1yt45dGD+cdXlW59//Kv13PtHRupC/j4a8bw8qMGVzN1qSZ0dGT2Wb9NEXEhpZWG93XOncAp7cQ3AGft5Zw5wJyO5KTqOvqSK+g94Nn/k92+ZQNbHl5C38GH7YqNmPpiRkx9MVAqZB/40ectZCXpEDiQvllS5/r6HzZwVEM9W5vbAPjdiq3ctGwLN/7dUdT37sVjW1sAuG/d0/z47s3cdNlRrH2ihbfMW8HN7z2aul7tLSkjaaeOrmb8LJn5I+CVhzYV1bqmX32XcS97E0T7/+N9/N4/MPzYF3VyVpLUM9g3S13Lms07+OV9TzDj+cN3xb79x8f5u5c0UN+79Cf4yEGlcaWb7n2C150wlPrevZg4vC+TRtSz+JGnqpK3VEs6Os34orLdXpSebedKwj3Y/d//FAQ0nPQKRp78CjYt/xN9Bg9nwKiJez1n4723cuSFl3dekpLUjdk3S13bP9+4hn981eFsbX7mrrkVG7Zz28NP8qmFa6nvHfzTuYdz8rgBPPrEDk4ZP2BXu8OH9ObRLTuqkbZUUzq6mvHryrZbgIeA6Yc8G9WEo//iI/QdNJwdT25h+ff/jfoRY3j0Dz9mysUf3Os5T655gF596unfML4TM5Wkbs2+WeqiFi7bwmEDe3Pi2P78fsXWXfGWtmTz06386G+ex58feYq/m7+KWy4/mmznY6i9THSTVKaj98z+ZaUTUe3oO6g0XabPwCEMPepUtjYtY/vm9dwz7/8AsP2Jx7nnWx/l2Ld+jD4DhwGw8d4/MMIpxpJ0yNg3S13XopXb+MWyLdx8/xM0tyRPNLfyvh+sYsyQPpx33BAigmnjB9Ar4PFtrYwZ0oc1m58ZiX10SwujB/ep4juQakOH7pmNiPERcV1ErIuItRHxg4hwiK0Hat3eTOv2p3ZtP/Hw3Qw4fDInXfYlTpj1GU6Y9Rn6Dh7BcW/7xK5CNrONjcv+yPBjX1jFzCWpe7FvlrquD59zOLf+w7H89v3H8MU3jufFkwfx+TdM4FXHDuF3K54E4MHHmtnRmowYUMc5xw7mx3dvprmljZUbt7Pi8Wamjetf5XchdX0dnWb8DeC7FI/RAd5axM6pRFLqulq2bebB678AQLa1Mvy40xk6+aR9nrN11TL6DB5B/bBRnZGiJPUU9s1SjbnklGF88PpHOOfL99OnLvjM68cTERw9qh+vPX4IZ3/pfnr3Cj752rGuZCx1QEeL2YbM/EbZ/jcj4vIK5KMurn7YKI6b+S/7bHPCrM88a3/wxOM49i0frWRaktQT2TdLNeD0yYM4ffIgAPr27sXn3zCh3Xbvefko3vNyP/iX9kdHH83zWES8NSLqiq+3AhsqmZgkSdon+2ZJUo/W0WL2r4BLgEeBNcAbAReekCSpeuybJUk9WkenGX8SmJmZGwEiYgTwaUodqSRJ6nz2zZKkHq2jxexJOztLgMx8PCJOqVBOkiTpue133xwRE4CrgcOBNmBuZn6+KISvBSZRel7tJWVF8mzgUqAVeG9m/qwC76UmtbVs575r/i/Z2kK2tTLs6Bcw9oyL2LZuJatu+iatO5rpO2Qkk1/7Lurq+9O8eT1LvzGbfsPHADBw7JFMPOcd1X0TklTDOlrM9oqI4bt9+tvRcyVJ0qF3IH1zC/APmfmniBgM3B4RNwHvABZm5pURcQVwBfDhiJgKzACOB8YCv4iIozOztULvqaZEXR+mXHIFdX37ka0tLPveHIZOPolVC7/NuDNnMHjCsTx2169Z+8efMPYlbwCgfugojpv5ySpnLkndQ0fvmf0M8LuI+GREfAL4HfBvlUtLkiQ9h/3umzNzTWb+qdh+ArgHGAdMB+YVzeYBFxbb04FrMrM5M1cAy4HTDvUbqVURQV3ffkDpcXXZ1goRPL1xDYPGHwPAkCOOZ9N9i6qZpiR1Wx0aXc3MqyNiEfBKIICLMnNpRTOTJEl7dbB9c0RMAk4BbgVGZ+aa4rprImLn80HGAX8oO62piKmQbW3c+62P0bxpLQ3TzmLgmCPpP3I8mx+4g2FHPZ+N9/2R7U88vqv99s3ruefq/0Nd3/6MfckbdhW9kqT91+GpwkUHaQErSVIXcaB9c0QMAn4AXJ6ZWyJir03b+7btXG8WMAtg4sSJ+5tOTYtevThu5idpefpJHrz+Czy1vokjzr2UVb/8Nmt+/yOGHXkKUVcHQJ+BwzjhnZ+jd/9BbHt0BQ9c/wWmvuP/Ulffv8rvQpJqk/e9SpLUg0REH0qF7Hcy84dFeG1EjClGZccA64p4EzCh7PTxwOrdr5mZc4G5AI2NjXsUuz1B734DGTzhWLY8dCejX/Aaplz8IQCefvxRNj/4ZwB69e5Dr959ABhw+GTqh47i6Y2PMvDwyVXLW5JqWUfvmZUkSTUuSkOwXwPuyczPlh1aAMwstmcC15fFZ0REfURMBqYAt3VWvl3djm1baHn6SQDadmxny8NL6TdiLDue3AJAZhuP/uF6Rp78yl3ts60NgOZN62je9Cj1Qxuqk7wkdQOOzEqS1HOcAbwNuCsiFhexfwSuBOZHxKXASuBigMxcEhHzKU1lbgEucyXjZ+x4chMP//S/SgVqJsOPOY2hR05j3e0/Z/3iXwAwbEojh53wUgC2Ni1jzW9/SPSqg+jFhHPeQe/+g6r5FiSpplnMSpLUQ2TmLbR/HyzAWXs5Zw4wp2JJ1bABDRM57u17PmZn1KmvYtSpr9ojPvzoFzD86Bd0RmqS1CM4zViSJEmSVHMsZiVJkiRJNcdiVpIkSZJUcypWzEbEhIj4VUTcExFLIuJ9RXxERNwUEfcXr8PLzpkdEcsjYllEnFup3CRJkiRJta2SI7MtwD9k5nHAi4DLImIqcAWwMDOnAAuLfYpjM4DjgfOAqyKiroL5SZIkSZJqVMWK2cxck5l/KrafAO4BxgHTgXlFs3nAhcX2dOCazGzOzBXAcuC0SuUnSZIkSapdnXLPbERMAk4BbgVGZ+YaKBW8wKii2ThgVdlpTUVs92vNiohFEbFo/fr1Fc1bkiRJktQ1Vfw5sxExCPgBcHlmbonY2+Pt2n3uXe4RyJwLzAVobGzc47gkdQVP72jjkm+sYHtL0tKWvGbqEP7+laO5bP5KHtywHYAtT7cypF8dP/3bo9i4rYV3XbuKO1c/xRunDeOTrx1b5XcgSZLUtVW0mI2IPpQK2e9k5g+L8NqIGJOZayJiDLCuiDcBE8pOHw+srmR+klQp9b2D782cxMD6Ona0Jm/82oOcOWUwX75k4q42n7xxDUP61RXte/GBV45i2bpmlq17ulppS5Ik1YxKrmYcwNeAezLzs2WHFgAzi+2ZwPVl8RkRUR8Rk4EpwG2Vyk+SKikiGFhfKlRbWpMdbUn5xJTM5H+WbOaCE4cCMKBvL15wxEDqe+919ookSZLKVHJk9gzgbcBdEbG4iP0jcCUwPyIuBVYCFwNk5pKImA8spbQS8mWZ2VrB/CSpolrbkvP/8wEeenw7b3/BCE4ZP2DXsdse3sbIQb2ZfFh9FTOUJEmqXRUrZjPzFtq/DxbgrL2cMweYU6mcJKkz1fUKfvq3R7H5qVZmXbOSZWuf5pjR/QBYcNdmLjhhWHUTlCRJqmGdspqxJPVkQ/vXcfqkgdy8fCtQmnZ84z2bed0JQ6ucmSRJUu2ymJWkCtjwZAubnyrdKfH0jjZueXArR43sC8AtD27lyJH1jBnap5opSpIk1bSKP5pHknqidU+08PfXNdGWSVvC+ccP5axjhgDw47s3c8GJw/Y454zPLeOJ5jZ2tCY/v3cL33rbJI4e1a+TM5ckSaoNFrOSVAHHHd6Pn/7tUe0e+8zrx7cb/+37j6lkSpIkSd2K04wlSZIkSTXHYlaSJEmSVHMsZiVJkiRJNcdiVpIkSZJUcyxmJUmSJEk1x2JWkiRJklRzLGYlSZIkSTXHYlaSJEmSVHMsZiVJkiRJNcdiVpIkSZJUcyxmJUmSJEk1x2JWkiRJklRzLGYlSeohIuLrEbEuIu4ui42IiJsi4v7idXjZsdkRsTwilkXEudXJWpKk9lnMSpLUc3wTOG+32BXAwsycAiws9omIqcAM4PjinKsioq7zUpUkad8sZiVJ6iEy89fA47uFpwPziu15wIVl8WsyszkzVwDLgdM6I09JkjrCYlaSpJ5tdGauASheRxXxccCqsnZNRWwPETErIhZFxKL169dXNFlJknaymJUkSe2JdmLZXsPMnJuZjZnZ2NDQUOG0JEkqsZiVJKlnWxsRYwCK13VFvAmYUNZuPLC6k3OTJGmvKlbMumKiJEk1YQEws9ieCVxfFp8REfURMRmYAtxWhfwkSWpXJUdmv4krJkqS1GVExPeA3wPHRERTRFwKXAmcExH3A+cU+2TmEmA+sBS4EbgsM1urk7kkSXvqXakLZ+avI2LSbuHpwJnF9jzgZuDDlK2YCKyIiJ0rJv6+UvlJktTTZOab93LorL20nwPMqVxGkiQduM6+Z9YVEyVJkiRJB62rLADliomSJEmSpA7r7GLWFRMlSZIkSQets4tZV0yUJEmSJB20ii0AVayYeCYwMiKagI9RWiFxfrF64krgYiitmBgRO1dMbMEVEyVJkiRJ+1DJ1YxdMVGSJEmSVBFdZQEoSZIkSZI6zGJWkiRJklRzKjbNWJIkSZIq5YwvnlHtFLSffvue3x7S6zkyK0mSJEmqORazkiRJkqSaYzErSZIkSao5FrOSJEmSpJpjMStJkiRJqjkWs5IkSZKkmuOjeSRJkiSpSlq2tbD8e8t5as1TEHDUXxzFxqUbefyuxyGgz6A+THnrFPoO7VvtVLsci1lJkiRJqpIVP1zB8OOGc+ylx9LW0kbb9jb6H96fia+dCMCa/13DqhtXceSbjqxypl2P04wlSZIkqQpanmphy/ItjDp9FAC9evei94De9O7/zJhja3NrtdLr8hyZlSRJkqQqaN7QTJ9BfVj+neVse2QbAycMZPIbJlNXX8fDNzzM+tvWU9e/jhPefUK1U+2SHJmVJEmSpCrItmRr01YOf8nhnPzhk6mrr+ORXzwCwBHnH0HjJxppOLWBNb9ZU+VMuyaLWUmSJEmqgr7D+lI/rJ7BkwYDcNi0w9i6auuz2oxsHMmGP2+oRnpdnsWsJEmSJFVB3yF96TusL0+tfQqATcs2MeDwATy17qldbTbetZH+o/pXK8UuzXtmJUmSJKlKnvfG53Hf1feRrUm/w/px1FuOKj2qZ91TRAT1w+t53pueV+00uySLWUmSJEmqkoHjB3LyB09+VuzYS4+tUja1xWnGkiRJkqSaYzErSZIkSao5FrOSJEmSpJpjMStJkiRJqjldrpiNiPMiYllELI+IK6qdjyRJPZ19sySpK+pSxWxE1AFfBl4NTAXeHBFTq5uVJEk9l32zJKmr6lLFLHAasDwzH8zM7cA1wPQq5yRJUk9m3yxJ6pK62nNmxwGryvabgBeWN4iIWcCsYndrRCzrpNzUcSOBx6qdRC2LT8+sdgrqHP6uHKyPxaG+4hGH+oLdgH3zwemWv+f2U4dEt/y3UYH/L/dE3fPfBhDvPaB/H3vtm7taMdveu8tn7WTOBeZ2Tjo6EBGxKDMbq52H1NX5u6IaYd98EPw91974b0N747+Njutq04ybgAll++OB1VXKRZIk2TdLkrqorlbM/hGYEhGTI6IvMANYUOWcJEnqyeybJUldUpeaZpyZLRHxbuBnQB3w9cxcUuW0tP+caiZ1jL8r6vLsmw+av+faG/9taG/8t9FBkZnP3UqSJEmSpC6kq00zliRJkiTpOVnMSpIkSZJqjsWsDomIyIj4TNn+ByLi41VMSeqSouSWiHh1WeySiLixmnlJOjgRMSki7t4t9vGI+EC1clJ1RURrRCwu+5oUEWdGxA3Pcd60iHhNZ+Wpyir7d7AkIv4cEX8fEdZgh4j/IXWoNAMXRcTIaicidWVZWqjgXf+/vXuPtquszz3+fRLCNYEkTcBISBMlIkEl1DReUEoFlbYoVgpGi0KlpQ6x1raocNrj9dAyrK16FGqpRaEqF0sV1B4BUQQFhACRe2oEhEAkhHswBpL8zh9rRhZhJ+zAXnvtufl+xlhjzfnO953rtzBx5ZnvvAD/nGTrJNsBxwNH97cySdIQW1VVc7tetw1y3FzAMDt6rP9zsAfwWjr/2364zzWNGoZZDZU1dO689lf9LkQa6arqeuCbwAfp/KCdVlU/629VknolyUVJPp3k0iTXJ5nf75rUf0nmN38mrmned2sef/Ux4C3NbN5b+l2nhk5VLQeOAt7TnKm1dZIvJrmu+XPwuwBJ/jvJS5rla5J8qFn+eJI/bWb4L0ryn0luTvKVJOnfN+ufEfVoHrXeicC1ST7R70KkFvgocDXwKDCvz7VI6r3tquqVSfYBTgFe1O+C1HPbJFnULN9aVX+4wfabgX2ax1/tD/x9VR3cBJd5VfWe4SxWw6OqbmlOM94ROKxpe3GSFwLnJ3kBcDHw6iS30Zkw2rsZ/irgy8A0YC9gD+Au4EdNnx8O41cZEQyzGjJV9VCS04D3Aqv6XY80klXVI0nOBFZW1ep+1yPpGdvYsw7Xt58OUFUXJ9k+ycSqemBYKlO/rKqquZvYvgNwapLZdP6cjBuWqjQSrJ9FfRXwWYCqujnJz4EXAJfQ+ff0rcC3gdcm2RaYWVWLk0wDrqiqpQDNQZOZPAvDrKcZa6h9GjgS2K7PdUhtsK55SWq/e4FJG7RNBlY0yxuG3Y2FXz17fBz4flW9CHgDsHWf69EwSPI8YC2wnMdD7YaupHPW1qvpzNJeA/wZcFVXn+4D4Wt5lk5SGmY1pKrqPuAsOoFWkqRnhapaCSxLsh9AksnAATw+U/KWpv1VwINV9WBfCtVIsgNwZ7N8RFf7w8CEYa9GPZdkKvB54HPNDSEvBv642fYCYAawuKoeBe4ADgUupzNTe0zzri6GWfXCPwHe1ViS9GzzDuDvmlP+vgd8tOvmbvcnuZTOP2Q94CuATwD/kORHwNiu9u8Dc7wB1KixzfpH8wDfBc6nc98MgJOAsUmuA84Ejui69OgS4O6q+mWzPB3D7JOkc1BAkiRJvZDkIuCYqlrY71okaTRxZlaSJEmS1DrOzEqSJEmSWseZWUmSJElS6xhmJUmSJEmtY5iVJEmSJLWOYVYaYknWNrdgX/86toef9bEk+z+D8eckuWwI6vhCkjnPdD+SJEnSYHkDKGmIJVlZVeOHcH+h83d13VDts9nvROA6YCXw+1V161DuX5Kk0SLJ3wJvA9YC64D7gUnAeGAqsP439N1VdekA4w8EPk5nImkc8Jmq+tckHwFWVtUne/4lpFHImVlpGCTZIcniJLs166cn+bNm+f1JrkxybZKPNm0zk9yU5CTgamCXJB9Icl2SnyQ5oen3pSR/1CyfkOTGZj+fbNqmJjm72f+VSfbuKutg4JvAGcCCrlq/lOT/Jrk0yS1d+x+T5KQkNyT5VpL/7tp2UZJ5zfLKJMc3dV6eZKem/Q1JfpzkmiTfXd8uSdJIluQVwIHAb1XVS4D9gT+uqrnAnwKXVNXc5jVQkB0HnAy8oar2BPYCLhqu+qXRzDArDb1tNjjN+C1V9SDwHuBLSRYAk6rq35K8DpgNzAfmAi9Nsk+zn92A06pqL2AO8CbgZc0P4Se6PzDJZOAPgT2aH9r/02z6DPCpqvptOuH1C13D3gqc3rzeusF3mAa8is6P9wlN25uBmcCL6fx4v2Ij33874PKmzouBP2vafwi8vPk+ZwAf2Mh4SZJGkmnAiqpaDVBVK6rqrs0YPwHYAri3Gb+6qhZv2CnJ3OYg8LVJvp5kUtN+UZJPNweZr08yv2nfLskpzcHqa5Ic9Ey/qNQ2W/S7AGkUWtUcrX2CqrogySHAicCeTfPrmtc1zfp4OuH2duDnVXV5074/8MWq+mWzr/s22P1DwK+ALyT5NvCtrnFzOmcqA7B9kgnAtsCuwA+rqpKsSfKiqrq+6feN5rTmG7tmUF8FfK1p/0WS72/k+z/a9flXAa9tlqcDZyaZBmzJ46dkSZI0kp0PfCjJ/wDfBc6sqh8MdnBV3ZfkXODnSS6k8xt5+gCXD50G/EVV/SDJx4APA+9rtm1XVa9sDnifArwI+Fvge1X1zubSoSuSfLeqHnn6X1VqF2dmpWGSZAywO7AKmLy+GfiHrtOTdq2qf2+2df8YBdjoBe5VtYbO7O7ZdGZwv9NsGgO8omv/O1fVw8Bb6Fzrc2uS2+jMuC7o2uXqDT67+/2pPFaPX4y/lscPmn0W+FxVvRj4c2DrQe5PkqS+qaqVwEuBo4B76ByYPWIz9/GnwH7AFcAxdALpryXZAZjYFZJPBfbp6nJ6s5+L6RyYnkjnYPixSRbROW15a2DG5tQltZ1hVho+fwXcROeU3lOaa2jOA96ZZDxAkp2T7DjA2PObfts2/SZ3b2zG71BV/03nKO7crnHv6eq3vv2twAFVNbOqZtL5ke4OswP5IXBwc+3sTsC+T/2Vn2AH4M5m+fDNHCtJUt9U1dqquqiqPkznd/Xgp7GP66rqU3TOWNrc8Rse0C46B5kP7jpgPaOqbtrcuqQ2M8xKQ2/Da2ZPSPICOteZ/k1VXULnWtK/q6rzga8ClyW5DvhPOtfWPEFVfQc4F1jYHIE9ZoMuE4BvJbkW+AGd4AzwXmBec/3NjcC7ksykc+R2/SnMNHcyfijJyzbxvc4GlgLXA/8K/Bh4cND/VeAjwNeSXAKs2IxxkiT1TZLdkszuapoL/Hwzxo9Psu+mxjf31rg/yaubprfT+T1f7y3Nvl4FPNj0Pw/4izTXEiXZa7A1SaOFj+aRNGhJxlfVyiS/QedUqb2r6hf9rkuSpF5J8lI6l8pMBNYAS4CjqmpFE1KPqaoDNzF+AnAm8Hw6lxo9AvxlVS1M16N5mrOnPk/nvha3AH9SVfcnuQi4DPgdYHvgnVV1RZJtgE8Dr6QzS3vbpuqQRiPDrKRBa35QJ9K5gdMnqupL/axHkqTRrvntPaaqFva7Fmmk8W7Gkgatqvbtdw2SJEkSODMrSZIkDYkkXwdmbdD8wao6rx/1SKOdYVaSJEmS1DrezViSJEmS1DqtvmZ2ypQpNXPmzH6XIUkaJa666qoVVTW133VIkqSn1uowO3PmTBYu9MZukqShkWTQz46UJEn95WnGkiRJkqTWMcxKkiRJklrHMCtJkiRJah3DrCRJkiSpdQyzkiRJkqTW6WmYTXJbkuuSLEqysGmbnOSCJD9t3id19T8uyZIki5O8vpe1SZIkSZLaazgezfO7VbWia/1Y4MKqOiHJsc36B5PMARYAewDPBb6b5AVVtXYYapS0mW7/2Iv7XYLEjA9d1+8SJElSn/TjNOODgFOb5VOBN3W1n1FVq6vqVmAJMH/4y5MkSZIkjXS9DrMFnJ/kqiRHNW07VdUygOZ9x6Z9Z+COrrFLmzZJkiRJkp6g16cZ711VdyXZEbggyc2b6JsB2upJnTqh+CiAGTNmDE2VkiRJkqRW6enMbFXd1bwvB75O57Thu5NMA2jelzfdlwK7dA2fDtw1wD5Prqp5VTVv6tSpvSxfkiRJkjRC9SzMJtkuyYT1y8DrgOuBc4HDm26HA+c0y+cCC5JslWQWMBu4olf1SZIkSZLaq5enGe8EfD3J+s/5alV9J8mVwFlJjgRuBw4BqKobkpwF3AisAY72TsaSJEmSpIH0LMxW1S3AngO03wvst5ExxwPH96omSZIkSdLo0I9H80iSJEmS9IwYZiVJkiRJrWOYlSRJkiS1jmFWkiRJktQ6hllJkiRJUusYZiVJkiRJrWOYlSRJkiS1jmFWkiRJktQ6hllJkiRJUusYZiVJkiRJrWOYlSRJkiS1jmFWkiRJktQ6hllJkiRJUusYZiVJkiRJrWOYlSRJkiS1jmFWkiRJktQ6hllJkiRJUusYZiVJkiRJrWOYlSRJkiS1jmFWkiRJktQ6hllJkiRJUusYZiVJkiRJrWOYlSRJkiS1jmFWkiRJktQ6hllJkiRJUuts0esPSDIWWAjcWVUHJpkMnAnMBG4DDq2q+5u+xwFHAmuB91bVeb2ub0Mvff9pw/2R0pNc9Y/v6HcJkiRJ0og2HDOzfwnc1LV+LHBhVc0GLmzWSTIHWADsARwAnNQEYUmSJEmSnqCnYTbJdOAPgC90NR8EnNosnwq8qav9jKpaXVW3AkuA+b2sT5IkSZLUTr2emf008AFgXVfbTlW1DKB537Fp3xm4o6vf0qbtCZIclWRhkoX33HNPT4qWJEmSJI1sPQuzSQ4EllfVVYMdMkBbPamh6uSqmldV86ZOnfqMapQkSZIktVMvbwC1N/DGJL8PbA1sn+TLwN1JplXVsiTTgOVN/6XALl3jpwN39bA+SZIkSVJL9WxmtqqOq6rpVTWTzo2dvldVhwHnAoc33Q4HzmmWzwUWJNkqySxgNnBFr+qTJEmSJLVXzx/NM4ATgLOSHAncDhwCUFU3JDkLuBFYAxxdVWv7UJ8kSZIkaYQbljBbVRcBFzXL9wL7baTf8cDxw1GTJEmSJKm9huM5s5IkSZIkDSnDrCRJkiSpdQyzkiRJkqTWMcxKkiRJklrHMCtJkiRJah3DrCRJkiSpdQyzkiRJkqTWMcxKkiRJklrHMCtJkiRJah3DrCRJkiSpdQyzkiRJkqTWMcxKkiRJklrHMCtJkiRJah3DrCRJkiSpdQyzkiRJkqTWMcxKkiRJklrHMCtJkiRJah3DrCRJkiSpdQyzkiRJkqTWMcxKkiRJklrHMCtJkiRJah3DrCRJkiSpdQyzkiRJkqTWMcxKkiRJklqnZ2E2ydZJrkjykyQ3JPlo0z45yQVJftq8T+oac1ySJUkWJ3l9r2qTJEmSJLVbL2dmVwOvqao9gbnAAUleDhwLXFhVs4ELm3WSzAEWAHsABwAnJRnbw/okSZIkSS3VszBbHSub1XHNq4CDgFOb9lOBNzXLBwFnVNXqqroVWALM71V9kiRJkqT2GlSYTXLhYNoG6DM2ySJgOXBBVf0Y2KmqlgE07zs23XcG7ugavrRp23CfRyVZmGThPffcM5jyJUmSJEmjzCbDbHPd62RgSpJJzfWuk5PMBJ77VDuvqrVVNReYDsxP8qJNfdxAuxhgnydX1byqmjd16tSnKkGSJEmSNApt8RTb/xx4H53gehWPB86HgBMH+yFV9UCSi+hcC3t3kmlVtSzJNDqzttCZid2la9h04K7BfoYkSZIk6dljkzOzVfWZqpoFHFNVz6uqWc1rz6r63KbGJpmaZGKzvA2wP3AzcC5weNPtcOCcZvlcYEGSrZLMAmYDVzzdLyZJkiRJGr2eamYWgKr6bJJXAjO7x1TVaZsYNg04tbkj8RjgrKr6VpLLgLOSHAncDhzS7OuGJGcBNwJrgKOrau3T+E6SJEmSpFFuUGE2yX8AzwcWAesDZgEbDbNVdS2w1wDt9wL7bWTM8cDxg6lJkiRJkvTsNagwC8wD5lTVk27IJEmSJEnScBvsc2avB57Ty0IkSZIkSRqswc7MTgFuTHIFsHp9Y1W9sSdVSZIkSZK0CYMNsx/pZRGSJEmSJG2Owd7N+Ae9LkSSJEmSpMEa7N2MH6Zz92KALYFxwCNVtX2vCpOkttv7U4vZbssxjB0Txo6Bb/35rjzwyzUc/bU7WPrAY0yfOI6TDp3BDtuM5bG1xQfPuZPrl61izTo4eM+JHL3P1H5/BUmSpBFrsDOzE7rXk7wJmN+LgiRpNDnjiFlM3u7x/6s96Ycr2Pt543n3q6dy0iX3cNIl93Dc657Dt294kEfXFucfPZtVj65j/xN/yhtfvAO7TNqyj9VLkiSNXIO9m/ETVNU3gNcMbSmSNPpdcPNDHDx3IgAHz53I+Tc/BEACv3x0HWvWFr9as45xY8OErZ7W/0VLkiQ9Kwz2NOM3d62OofPcWZ85K0lP4bD/uI0AfzxvMm+bN5kVj6xhpwnjANhpwjhWPLIGgN+fswMX3Pwwv/3Jm1n12Do+dMA0Jm472Hv0SZIkPfsM9l9Kb+haXgPcBhw05NVI0ijyX0c+j522H8eKlWs47LTbeP6UrTbad9Gdv2RM4IpjXsiDq9ZyyCm38KrnjWfGZE8zliRJGshgr5n9k14XIkmjzU7bd2Zgp4zfgtfvPoFFd65iynZbcPfDj7HThHHc/fBjTGmupz3n2gfZd/Z4xo0NU8ZvwUtnbMu1d60yzEqSJG3EoC7ISjI9ydeTLE9yd5Kzk0zvdXGS1Fa/fHQdK1ev/fXyxT9byW47bsX+u23P2YseAODsRQ/w2hd2bgq/8w7juPSWR6gqfvnoOq5ZumqTM7mSJEnPdoM9zfiLwFeBQ5r1w5q21/aiKElquxUr13DUGbcDsGZdcdCLd2Df2RPYc+dtePdZd3Dm1ffz3B3G8S+H7gLAO+ZP5phv3MlrT1xCAYfMncjuz9m6j99AkiRpZBtsmJ1aVV/sWv9Skvf1oB5JGhVmTN6S77x71ye1T9p2C04/YtaT2rfbaiz/8pYZw1GaJEnSqDDY5z6sSHJYkrHN6zDg3l4WJkmSJEnSxgw2zL4TOBT4BbAM+CPAm0JJkiRJkvpisKcZfxw4vKruB0gyGfgknZArSZIkSdKwGuzM7EvWB1mAqroP2Ks3JUmSJEmStGmDDbNjkkxav9LMzA52VleSJEmSpCE12ED6T8ClSf4TKDrXzx7fs6okSZIkSdqEQYXZqjotyULgNUCAN1fVjT2tTJIkSZKkjRj0qcJNeDXACoBat46bv/xhxo2fxK5v/mvuX3wFyy79Or+6dxm7HfZhtnvO48/R/MWPv8m9110MGcMurzmM7We9uI+VS5IkSRoNBnvNrPQEy68+n60nP/fX61tPmc7zDnov46fv9oR+q1bcyf03/5jdj/h7dj34GG7/7qnUunXDXa4kSZKkUcYwq8326MP38dAtP2HKS37n123b/MZz2XrytCf1ffBnVzPphS9jzBbj2GriVLaatBOP/OKW4SxXkiRJ0ijUszCbZJck309yU5Ibkvxl0z45yQVJftq8d98l+bgkS5IsTvL6XtWmZ2bp977CzvscSufy6U177OH7GTdh8q/Xtxw/mccevn8TIyRJkiTpqfVyZnYN8DdVtTvwcuDoJHOAY4ELq2o2cGGzTrNtAbAHcABwUpKxPaxPT8ODP1vEFttuz7Zd18RuSg3U+NQZWJIkSZI2qWfPiq2qZcCyZvnhJDcBOwMHAfs23U4FLgI+2LSfUVWrgVuTLAHmA5f1qkZtvpV3/g8P/uwarr/1WtateYy1j67i1m9/nll/8K4B+285YRKPPXzfr9cfXXkf48ZPGrCvJEmSJA1Wz8JstyQzgb2AHwM7NUGXqlqWZMem287A5V3DljZtG+7rKOAogBkzZvSwag1k530ObU4xhodvv4m7F/6/jQZZgB2evxe3ffvz7PjSA3hs5QOsvv9utnvO84arXEmSJEmjVM/DbJLxwNnA+6rqoWSj55gOtOFJZ6lW1cnAyQDz5s0b8CxWDb8HfrqQOy78MmtWPczP/uuf2WbHGcz+o/ezzZTpTNxtPjd+8TgyZiy77Pd2Msb7jkmSJEl6ZnoaZpOMoxNkv1JV/9U0351kWjMrOw1Y3rQvBXbpGj4duKuX9emZmTBjdybM2B2AibPnMXH2vAH7TXv5G5n28jcOZ2mSJEmSRrle3s04wL8DN1XVP3dtOhc4vFk+HDinq31Bkq2SzAJmA1f0qj5JkiRJUnv1cmZ2b+DtwHVJFjVt/ws4ATgryZHA7cAhAFV1Q5KzgBvp3An56Kpa28P6JEmSJEkt1cu7Gf+QjT+EZb+NjDkeOL5XNUmSJEmSRgfvxCNJkiRJah3DrCRJkiSpdQyzkiRJkqTWMcxKkiRJklrHMCtJkiRJah3DrCRJkiSpdQyzkiRJkqTWMcxKkiRJklrHMCtJkiRJah3DrCRJkiSpdQyzkiRJkqTWMcxKkiRJklrHMCtJkiRJah3DrCRJkiSpdQyzkiRJkqTWMcxKkiRJklrHMCtJkiRJah3DrCRJkiSpdQyzkiRJkqTWMcxKkiRJklrHMCtJkiRJah3DrCRJkiSpdQyzkiRJkqTWMcxKkiRJklqnZ2E2ySlJlie5vqttcpILkvy0eZ/Ute24JEuSLE7y+l7VJUmSJElqv17OzH4JOGCDtmOBC6tqNnBhs06SOcACYI9mzElJxvawNkmSJElSi/UszFbVxcB9GzQfBJzaLJ8KvKmr/YyqWl1VtwJLgPm9qk2SJEmS1G7Dfc3sTlW1DKB537Fp3xm4o6vf0qZNkiRJkqQnGSk3gMoAbTVgx+SoJAuTLLznnnt6XJYkSZIkaSQa7jB7d5JpAM378qZ9KbBLV7/pwF0D7aCqTq6qeVU1b+rUqT0tVpIkSZI0Mg13mD0XOLxZPhw4p6t9QZKtkswCZgNXDHNtkiRJkqSW2KJXO05yOrAvMCXJUuDDwAnAWUmOBG4HDgGoqhuSnAXcCKwBjq6qtb2qTZIkSZLUbj0Ls1X11o1s2m8j/Y8Hju9VPZIkSZKk0WOk3ABKkiRJkqRBM8xKkiRJklrHMCtJkiRJah3DrCRJkiSpdQyzkiRJkqTWMcxKkiRJklrHMCtJkiRJah3DrCRJkiSpdQyzkiRJkqTWMcxKkiRJklrHMCtJkiRJah3DrCRJkiSpdQyzkiRJkqTWMcxKkiRJklrHMCtJkiRJah3DrCRJkiSpdQyzkiRJkqTWMcxKkiRJklrHMCtJkiRJah3DrCRJkiSpdQyzkiRJkqTWMcxKkiRJklrHMCtJkiRJah3DrCRJkiSpdQyzkiRJkqTWGXFhNskBSRYnWZLk2H7XI0mSJEkaeUZUmE0yFjgR+D1gDvDWJHP6W5UkSZIkaaQZUWEWmA8sqapbqupR4AzgoD7XJEmSJEkaYbbodwEb2Bm4o2t9KfCy7g5JjgKOalZXJlk8TLVp8KYAK/pdRJvlk4f3uwQND/+uPFMfzlDv8TeHeoeSJKk3RlqYHehfJfWElaqTgZOHpxw9HUkWVtW8ftchjXT+XZEkSXr6RtppxkuBXbrWpwN39akWSZIkSdIINdLC7JXA7CSzkmwJLADO7XNNkiRJkqQRZkSdZlxVa5K8BzgPGAucUlU39LksbT5PA5cGx78rkiRJT1Oq6ql7SZIkSZI0goy004wlSZIkSXpKhllJkiRJUusYZjVkkhyQZHGSJUmO7Xc90kiV5JQky5Nc3+9aJEmS2sowqyGRZCxwIvB7wBzgrUnm9LcqacT6EnBAv4uQJElqM8Oshsp8YElV3VJVjwJnAAf1uSZpRKqqi4H7+l2HJElSmxlmNVR2Bu7oWl/atEmSJEnSkDPMaqhkgDaf+yRJkiSpJwyzGipLgV261qcDd/WpFkmSJEmjnGFWQ+VKYHaSWUm2BBYA5/a5JkmSJEmjlGFWQ6Kq1gDvAc4DbgLOqqob+luVNDIlOR24DNgtydIkR/a7JkmSpLZJlZc1SpIkSZLaxZlZSZIkSVLrGGYlSZIkSa1jmJUkSZIktY5hVpIkSZLUOoZZSZIkSVLrGGalYZBk5QbrRyT53BDte2aSt3Wt75vkwSTXJFmc5OIkB3Ztf1eSdwzFZ0uSJEn9skW/C5D09CXZApgJvA34atemS6rqwKbPXOAbSVZV1YVV9flhL1SSJEkaYs7MSn2WZGqSs5Nc2bz2btrnJ7m0mWG9NMluTfsRSb6W5JvA+cAJwKuTLEryVxvuv6oWAR8D3tOM/0iSY5rl9ya5Mcm1Sc5o2rZLckpTyzVJDmraZya5JMnVzeuVTfu0ZvZ3UZLrk7y6aX9dksuavl9LMr63/yUlSZL0bOLMrDQ8tkmyqGt9MnBus/wZ4FNV9cMkM4DzgN2Bm4F9qmpNkv2BvwcObsa8AnhJVd2XZF/gmK6Z2H0H+PyrgfcP0H4sMKuqVieZ2LT9LfC9qnpn03ZFku8Cy4HXVtWvkswGTgfm0ZkVPq+qjk8yFtg2yRTg74D9q+qRJB8E/ppOqJYkSZKeMcOsNDxWVdXc9StJjqATBAH2B+YkWb95+yQTgB2AU5vgWMC4rv1dUFX3bcbnZyPt1wJfSfIN4BtN2+uAN66fvQW2BmYAdwGfa05bXgu8oNl+JXBKknHAN6pqUZLfAeYAP2q+15bAZZtRryRJkrRJhlmp/8YAr6iqVd2NST4LfL+q/jDJTOCirs2PbOZn7AXcNED7HwD7AG8E/neSPegE34OravEG9XwEuBvYs6n5VwBVdXGSfZp9/UeSfwTupxO437qZdUqSJEmD4jWzUv+dT3M9K/z6hk3QmZm9s1k+YhPjHwYmbGxjkpcA/xs4cYP2McAuVfV94APARGA8ndOc/yLNlGqSvbrqWVZV64C3A2Ob7b8JLK+qfwP+Hfgt4HJg7yS7Nn22TbJ+JleSJEl6xgyzUv+9F5jX3ITpRuBdTfsngH9I8iOa4LgR1wJrkvyk6wZQr17/aB46Ifa9VXXhBuPGAl9Och1wDZ3rdh8APk7nlOZrk1zfrAOcBBye5HI6pxivnx3eF1iU5Bo61/R+pqruoRPAT09yLZ1w+8LN+q8iSZIkbUKqqt81SJIkSZK0WZyZlSRJkiS1jmFWkiRJktQ6hllJkiRJUusYZiVJkiRJrWOYlSRJkiS1jmFWkiRJktQ6hllJkiRJUuv8fw2lOI9072QxAAAAAElFTkSuQmCC\n",
"text/plain": [
"
"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"categorical_cols = [\"Sex\", \"ChestPainType\", \"FastingBS\", \"RestingECG\", \"ExerciseAngina\", \"ST_Slope\", \"HeartDisease\"]\n",
"\n",
"fig = plt.figure(figsize=(16,15))\n",
"\n",
"for idx, col in enumerate(categorical_cols):\n",
" ax = plt.subplot(4, 2, idx+1)\n",
" sns.countplot(x=df[col], ax=ax)\n",
" # add data labels to each bar\n",
" for container in ax.containers:\n",
" ax.bar_label(container, label_type=\"center\")"
]
},
{
"cell_type": "markdown",
"id": "2d97efb3",
"metadata": {},
"source": [
"- The dataset is highly skewed towards male patients. There are `725` male patients and `193` female patients. This could potentially induce a bias in our model.\n",
"- `496` patients had `ASY` (asymptotic) chest pain type.\n",
"- `552` patients had a normal restin ECG.\n",
"- `704` patients had blood sugar lower than `120` mg/dl\n",
"\n",
"Grouping these by `HeartDisease` will give us a better idea about the data distribution."
]
},
{
"cell_type": "code",
"execution_count": 10,
"id": "a5f28e00",
"metadata": {},
"outputs": [
{
"data": {
"image/png": "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\n",
"text/plain": [
"
"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"fig = plt.figure(figsize=(16,15))\n",
"\n",
"for idx, col in enumerate(categorical_cols[:-1]):\n",
" ax = plt.subplot(4, 2, idx+1)\n",
" # group by HeartDisease\n",
" sns.countplot(x=df[col], hue=df[\"HeartDisease\"], ax=ax)\n",
" # add data labels to each bar\n",
" for container in ax.containers:\n",
" ax.bar_label(container, label_type=\"center\")"
]
},
{
"cell_type": "markdown",
"id": "0b44da4c",
"metadata": {},
"source": [
"- We can further notice how skewed the dataset is towards male patients. Only `50` female patients in the dataset have been diagnosed with heart disease.\n",
"- A significant number of patients, `392`, diagnosed with heart disease have asymptomatic (ASY) chest pain. While chest pain could be a relevant feature for our model, asymptomatic implies that those patients who had a heart disease did not have chest pain as a symptom. \n",
"- A high number (`170`) of patients with blood sugar greater than 120 mg/dl were diagnosed with heart disease in relation to those who were not diagnosed as such.\n",
"- Out of all patients who had an exercise-induced angina, `316` were diagnosed with a heart disease.\n",
"- Out of all patients with a flat ST slope, `381` were diagnosed with a heart disease.\n",
"\n",
"Looking at the data distribution from the above plots, we can start to identify some features that could be relevant to us. We will clean up the dataset a bit first before narrowing down on our features."
]
},
{
"cell_type": "markdown",
"id": "9bb5c3c5",
"metadata": {},
"source": [
"## Data Cleaning\n",
"\n",
"We identified that there are no missing values. However, as we noticed earlier, a couple of columns have 0 values which don't make sense.\n",
"\n",
"We will look at how many zero values `RestingBP` and `Cholesterol` contain and decide how to handle those."
]
},
{
"cell_type": "code",
"execution_count": 11,
"id": "3ad28327",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"
"
],
"text/plain": [
" Age Sex ChestPainType RestingBP Cholesterol FastingBS RestingECG \\\n",
"293 65 M ASY 115 0 0 Normal \n",
"294 32 M TA 95 0 1 Normal \n",
"295 61 M ASY 105 0 1 Normal \n",
"296 50 M ASY 145 0 1 Normal \n",
"297 57 M ASY 110 0 1 ST \n",
".. ... .. ... ... ... ... ... \n",
"514 43 M ASY 122 0 0 Normal \n",
"515 63 M NAP 130 0 1 ST \n",
"518 48 M NAP 102 0 1 ST \n",
"535 56 M ASY 130 0 0 LVH \n",
"536 62 M NAP 133 0 1 ST \n",
"\n",
" MaxHR ExerciseAngina Oldpeak ST_Slope HeartDisease \n",
"293 93 Y 0.0 Flat 1 \n",
"294 127 N 0.7 Up 1 \n",
"295 110 Y 1.5 Up 1 \n",
"296 139 Y 0.7 Flat 1 \n",
"297 131 Y 1.4 Up 1 \n",
".. ... ... ... ... ... \n",
"514 120 N 0.5 Up 1 \n",
"515 160 N 3.0 Flat 0 \n",
"518 110 Y 1.0 Down 1 \n",
"535 122 Y 1.0 Flat 1 \n",
"536 119 Y 1.2 Flat 1 \n",
"\n",
"[172 rows x 12 columns]"
]
},
"execution_count": 12,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df[df[\"Cholesterol\"] == 0]"
]
},
{
"cell_type": "markdown",
"id": "1dbe3eca",
"metadata": {},
"source": [
"`RestingBP` has only one zero value. We can remove that row from consideration. There are `172` zero values for `Cholesterol`. That's a relatively high number. We can't remove them all and replacing those values with the median might not be an ideal approach, but that's what we will go for now.\n",
"\n",
"To be more accurate, we will replace the zero values in `Cholesterol` in relation to `HeartDisease`. So, the 0 values in `Cholesterol` for patients who were diagnosed with a heart disease will be replaced by the median of the non-zero values for patients who were diagnosed with a heart disase. And we'll do the same for those who were not diagnosed with a heart disease."
]
},
{
"cell_type": "code",
"execution_count": 13,
"id": "61c9e24a",
"metadata": {},
"outputs": [],
"source": [
"df_clean = df.copy()\n",
"\n",
"# only keep non-zero values for RestingBP\n",
"df_clean = df_clean[df_clean[\"RestingBP\"] != 0]\n",
"\n",
"heartdisease_mask = df_clean[\"HeartDisease\"]==0\n",
"\n",
"cholesterol_without_heartdisease = df_clean.loc[heartdisease_mask, \"Cholesterol\"]\n",
"cholesterol_with_heartdisease = df_clean.loc[~heartdisease_mask, \"Cholesterol\"]\n",
"\n",
"df_clean.loc[heartdisease_mask, \"Cholesterol\"] = cholesterol_without_heartdisease.replace(to_replace = 0, value = cholesterol_without_heartdisease.median())\n",
"df_clean.loc[~heartdisease_mask, \"Cholesterol\"] = cholesterol_with_heartdisease.replace(to_replace = 0, value = cholesterol_with_heartdisease.median())"
]
},
{
"cell_type": "code",
"execution_count": 14,
"id": "6408d9a3",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"
\n",
"\n",
"
\n",
" \n",
"
\n",
"
\n",
"
Cholesterol
\n",
"
RestingBP
\n",
"
\n",
" \n",
" \n",
"
\n",
"
count
\n",
"
917.000000
\n",
"
917.000000
\n",
"
\n",
"
\n",
"
mean
\n",
"
239.700109
\n",
"
132.540894
\n",
"
\n",
"
\n",
"
std
\n",
"
54.352727
\n",
"
17.999749
\n",
"
\n",
"
\n",
"
min
\n",
"
85.000000
\n",
"
80.000000
\n",
"
\n",
"
\n",
"
25%
\n",
"
214.000000
\n",
"
120.000000
\n",
"
\n",
"
\n",
"
50%
\n",
"
225.000000
\n",
"
130.000000
\n",
"
\n",
"
\n",
"
75%
\n",
"
267.000000
\n",
"
140.000000
\n",
"
\n",
"
\n",
"
max
\n",
"
603.000000
\n",
"
200.000000
\n",
"
\n",
" \n",
"
\n",
"
"
],
"text/plain": [
" Cholesterol RestingBP\n",
"count 917.000000 917.000000\n",
"mean 239.700109 132.540894\n",
"std 54.352727 17.999749\n",
"min 85.000000 80.000000\n",
"25% 214.000000 120.000000\n",
"50% 225.000000 130.000000\n",
"75% 267.000000 140.000000\n",
"max 603.000000 200.000000"
]
},
"execution_count": 14,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"df_clean[[\"Cholesterol\", \"RestingBP\"]].describe()"
]
},
{
"cell_type": "markdown",
"id": "512c073a",
"metadata": {},
"source": [
"The minimum values for both have changed! There are no more zero values in either of those."
]
},
{
"cell_type": "markdown",
"id": "2c71332d",
"metadata": {},
"source": [
"## Feature Selection\n",
"\n",
"Thanks to our EDA and a general understanding of the features, we can identify some of the features that we could start with:\n",
"\n",
"- `Age`\n",
"- `Sex`\n",
"- `ChestPainType`\n",
"- `Cholesterol`\n",
"- `FastingBS`\n",
"\n",
"\n",
"We will also identify how stronly the feature columns are correlated to the target colummn. That should help us narrow down on the features.\n",
"\n",
"In order to do that, we will first convert our categorical columns into dummy variables."
]
},
{
"cell_type": "code",
"execution_count": 15,
"id": "22829bb7",
"metadata": {},
"outputs": [
{
"data": {
"text/html": [
"