Unverified Commit c3c92e21 authored by Simon Bowly's avatar Simon Bowly
Browse files

Week 2 activity

parent 5ceae43e
%% Cell type:markdown id:a3948254 tags:
# ADS1002 Week 2: Minimising errors to fit models
%% Cell type:code id:ce0cb9ab tags:
``` python
# Load the same datasets as in the introductory notebook.
import pandas as pd
import seaborn as sns
import numpy as np
import matplotlib.pyplot as plt
# Note: you can read data files directly from the internet
# using pd.read_csv(...)
base = "https://gitlab.erc.monash.edu.au/bads/data-challenges-resources/-/raw/main/Machine-Learning/Supervised-Methods/"
who_data_2015 = (
pd.read_csv(base + "who-health-data.csv") # Read in the csv data.
.rename(columns=lambda c: c.strip()) # Clean up column names.
.query("Year == 2015") # Restrict the dataset to records from 2015.
)
wisconsin_cancer_biopsies = (
pd.read_csv(base + "kaggle-wisconsin-cancer.csv")
.assign(diagnosis=lambda df: df['diagnosis']
.map({"M": "malignant", "B": "benign"})
.astype('category')
)
)
```
%% Cell type:markdown id:8c12f566 tags:
## Minimising errors (in two parameters) on the regression dataset
The code below produces colour maps to explore the variation of mean absolute error (MAE) using a linear predictive model.
%% Cell type:code id:81c32b7d tags:
``` python
# Don't worry too much about the mechanics of the code in this cell; it computes
# model predictive error for various parameters so that we can produce a general
# plot to guide choices of the fitted parameters.
def prediction_error(gradient, intercept):
""" Return the prediction error associated with the value of the parameters. """
who_with_predictions = (
who_data_2015[["Schooling", "Life expectancy"]]
# Add a column for our predictions.
.assign(Predicted=lambda df: df["Schooling"] * gradient + intercept)
.dropna()
)
errors = who_with_predictions['Life expectancy'] - who_with_predictions['Predicted']
return errors.abs().mean()
gradient_values, intercept_values = np.meshgrid(
np.linspace(1.0, 3.0, 30),
np.linspace(35, 80, 30),
)
errors = np.zeros(gradient_values.shape)
for i in range(errors.shape[0]):
for j in range(errors.shape[1]):
errors[i, j] = prediction_error(gradient_values[i, j], intercept_values[i, j])
```
%% Cell type:markdown id:60e46eca tags:
The cell below allows you to choose the gradient and intercept of a model in the form
$$
\text{Life expectancy} = \text{gradient} \times \text{Schooling} + \text{intercept}.
$$
This is a very general linear model. Certain choices of the model lead to low errors, and others to high errors.
Try changing the `gradient` and `intercept` values, and re-run the cell to examine the results. Then discuss the following:
* Is it easy to determine the parameters which give the smallest possible error? Is there an 'obvious' best result?
* Can you comment on the shape of this 2D error function?
* Are there any techniques you can think of (which you may have learned in previous mathematics courses) which might
%% Cell type:code id:b2eee8b5 tags:
``` python
# SET THESE VALUES AND TEST THE OUTPUT.
gradient = 1.5
intercept = 66
# This code plots the coloured background showing how mean absolute error
# changes with these two parameters.
plt.contourf(gradient_values, intercept_values, errors)
plt.xlabel("Gradient")
plt.ylabel("Intercept")
plt.colorbar(label="Error")
# Plots our chosen parameters on the colour map as a red point.
plt.scatter(gradient, intercept, c='r')
# Generate predictions using the selected gradient and intercept.
who_with_predictions = (
who_data_2015[["Schooling", "Life expectancy"]]
# Add a column with a computed prediction based on years of schooling.
.assign(Predicted=lambda df: df["Schooling"] * gradient + intercept)
# Discard for the moment any row where we can't make a prediction
# due to missing data.
.dropna()
)
# Plot both the predicted and actual life expectancy results against years of schooling.
sns.relplot(data=who_with_predictions.set_index("Schooling"));
# Also display the prediction error.
prediction_error(gradient, intercept)
```
%%%% Output: execute_result
13.678034682080925
%%%% Output: display_data
![](data:image/png;base64,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)
%%%% Output: display_data
![](data:image/png;base64,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)
%% Cell type:markdown id:cae5a84b tags:
## How errors change with different datasets
The cell below plots the error function of the classification model given different values of the `radius` split parameter (as in the introductory notebook). Each time this cell is run it will select two different subsets of the data, and plot the accuracy function (number of correct predictions) for each set. Investigate how the plots vary when you change the value of the parameter `N` (which controls the size of the sampled data used to calculate the accuracy curves), and when you re-run the cell for the same value of `N`. What do you observe in general?
Questions to discuss:
* Do the two curves (in blue and orange) generally have the same or different peaks?
* Does the location (parameter value) of the peak vary at all with the sample size?
* Does the best possible accuracy vary much?
%% Cell type:code id:33c49166 tags:
``` python
# VARY THIS PARAMETER BETWEEN 10 and 200
N = 10
def model_correct_predictions(data_subset, radius_split_parameter):
""" Return the number of correct predictions made by the model
for the given parameter value. """
data = data_subset.assign(
predicted=lambda df: df['radius_mean'].lt(radius_split_parameter)
.map({True: "benign", False: "malignant"})
)
return (data['diagnosis'] == data['predicted']).sum()
# Select two different random subsets of the data, given the size parameter.
train = wisconsin_cancer_biopsies.sample(N)
test = wisconsin_cancer_biopsies.sample(N)
test_parameter_values = np.arange(0, 30, 0.5) # Values spread across the interval 0 -> 30, separated by 0.5
cost_function_results = pd.DataFrame({
"Parameter Value": test_parameter_values,
# Use a Python list comprehension to find MAE for each proposed parameter value.
"Number of Correct Predictions (Train)": [model_correct_predictions(train, m) for m in test_parameter_values],
"Number of Correct Predictions (Test)": [model_correct_predictions(test, m) for m in test_parameter_values],
})
sns.relplot(data=cost_function_results.set_index("Parameter Value"), kind='line');
```
%%%% Output: display_data
![](data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAjEAAAFuCAYAAABwX0lLAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjQuMiwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8rg+JYAAAACXBIWXMAAAsTAAALEwEAmpwYAABB3UlEQVR4nO3deXzcdbX/8dfJOmmTTGhL20lL2ZciOwEFZEevC4L7VfECXrnIvV4EFRV/ehXR6wVX3FC5LtcFFwQR3BBEdtkKlLbsyNok3SiZpG2WJjm/P77faSfJJJlklu/M5P18PPLILN/5fk8mhTk5n8/nfMzdERERESk3VVEHICIiIjIdSmJERESkLCmJERERkbKkJEZERETKkpIYERERKUs1xbzY6173Or/hhhuKeUkRkXyyqAMQke2KWonZsGFDMS8nIiIiFUzDSSIiIlKWlMSIiIhIWVISIyIiImVJSYyIiIiUJSUxIiIiUpaUxIiIiEhZUhIjIiIiZUlJjIiIiJQlJTEiIiJSlpTEiIiISFmaNIkxsx+Z2TozW5X22Bwzu8nMngq/71DYMEVERERGyqYS83/A60Y9diFws7vvCdwc3hcREREpmkl3sXb3281sl1EPnwocF97+CXAr8Il8BiaSd2tWwQ1p+faux8CxH48uHhERycl058QscPfO8PYaYMF4B5rZ2Wa2zMyWrV+/fpqXE8nBlo2w+SXAYXgo+HrpH3DP5VFHJiIiOch5Yq+7O+ATPH+Fu7e5e9uOO+6Y6+VEpu7u78BX9oAdl8K//jn4Ouz90PsyDGyJOjoREZmm6SYxa80sARB+X5e/kETyrLsDmhJQnTZ6uugQOOBdMNgXXVwiIpKT6SYx1wNnhLfPAK7LTzgiBdC9GpoXjXxs9xPgrd+HWXOiiUlERHKWzRLrXwJ3A3ub2Wozez9wCfAaM3sKOCm8L1Kaku3Q3DryMffg8c0vRROTiIjkLJvVSe8e56kT8xyLSP65Q3c77P36kY8P9sPX94XjPw3Hfiya2EREJCfq2CuVbWATxBfDnF1HPl4bg1nzgqEmEREpS5NWYkTKWn0TnPtA5ufii4IhJRERKUuqxMjM1bw4GGoSEZGypCRGKtuyH8PX9g0a3o2mSoyISFlTEiOVret52LQWYvGxz83bK5gvM9hf/LhERCRnmhMjla27A5paoap67HOH/1vwJSIiZUmVGKlsyfZg2GgiPu6uGSIiUsKUxEhly9StN6VnDXxpN3joZ8WNSURE8kJJjFQud+juHNutN2XW3GDCryb3ioiUJc2JkcplBp94FoYHMz9fXQuNC9TwTkSkTCmJkcpWN3vi55tbVYkRESlTGk6SyvXsHfCLf4bkBJWW+CI1vBMRKVNKYqRyrXsUnrwBquvHP6Z5cTBvRiuURETKjpIYqVzJ1UECM3ve+Mec8Olg3oxZ8eISEZG80JwYqVzd7cGcl4kSlPrG4sUjIiJ5pUqMVK7ujmBbgYlsfBZ+eio8d1dxYhIRkbxREiOVK9k+fqO7lOo6eOZW2PBEUUISEZH80XCSVK63/xBqZ018TNNCsGotsxYRKUNKYqRy7XT45MdUVQeJTHdH4eMREZG80nCSVKaNz8DfvgBdL05+bPMide0VESlDSmKkMnWugNu/DH1dkx8bX6ThJBGRMqThJKlMqS68k03sBTjm4+PvryQiIiVLSYxUpmQ71DRAww6TH7tg38LHIyIieafhJKlM3auDYaJsOvFufAZu/C94+bmChyUiIvmjJEYqUzY9YlK2vAx//yasfbSwMYmISF5pOEkq02FnQW0su2PjYbKj3axFRMqKkhipTAe9O/tjZ8+HqlolMSIiZUbDSVJ5+pLw6PWwaX12x1dVQXNCy6xFRMqMkhipPOseg6v+BTqXZ/+a5sWqxIiIlBkNJ0nlmUqPmJRXnQPuhYlHREQKQkmMVJ7UsFB8CknMvqcWJhYRESkYDSdJ5eluh7pGqG+ewms6YMVvoL+ncHGJiEheKYmRypNcHQwlZdPoLqXjIfjtWbDhqcLFJSIieZVTEmNm55nZKjN7xMzOz1NMIrlpPRiWnjy11zSrV4yISLmZ9pwYM9sP+DfgcGAAuMHM/uDuT+crOJFpOeaCqb8mlcRombWISNnIZWLvUuBed98CYGa3AW8FvpSPwKQ8nfLtO1mxOpnzeZrZxC31H2WujZ2jMvC2n1K3/6lwz3fhhgvHvviV58DrL53aBWfPCzaMvOET8PRf4b1Xw8AW+GJi7LE1DfDpNcHtK46Hjgfh6I/CiZ+Z2jVFRCQnuSQxq4D/NrO5QC/wBmDZ6IPM7GzgbIAlS5bkcDkpdb0DQ6xYneToPedxyJIsdo+exDX9v+KgNb9hyGoB2Li5n6fWbeKtdUtYArCoDY7NkMQsbpv6xczgrVfA2kdg7u7BY9W1mc9fnfafzaFnQO9GeP7uqV9TRERyMu0kxt0fM7NLgRuBzcByYCjDcVcAVwC0tbWpEUcF60z2AvDmgxbxtkMX5+GMewGv3nbv7n+8xGX/ew9tVTsFScxOhwVf+bLvKcFXSnUtHP/JiV9z6Jnw3J3w4r35i0NERLKS08Red/+hux/q7scALwNP5icsKUedyT4AEi1Zbrw4keW/hCvfCYMD2x5qDc/bESZLJaN5EXR3wvBw1JGIiMwoua5Omh9+X0IwH+YX+QhKylMqiWmNN+R+so6H4IW7oaZu20MLmoMkZk14nZIRXww+BFs2RB2JiMiMkmvH3mvCOTFbgQ+6e1fuIUm56uwKKiQL43moxHS3j9k2IFZbzdzZdduGrUrGIafDoe8bOVdGREQKLqf/67r70fkKRMpfR7KPObPriNVW536y7vaM2wYkWmJ0dJVYJaamPuoIRERmJHXslbzpTPaSyEcVBoJ+Lc2tYx5OxBtKrxIzsBl+cgo8/OuoIxERmVGUxEjedHb1kcjHfJjBfti8DprHrnBqjcfoLLVKTO2sYHXS2pVRRyIiMqNoEF/ypiPZyyt3m5P7iawa3n8TNM4f81SipYGe/kF6+rbSFKvN/Vr5YBZUjdTtV0SkqJTESF5s6h+kp28wP5WY6hrY6fCMT6WGq9Yk+0oniYFwmbWSGBGRYtJwkuTFmnCeSms+esS8cC/cfDH0jd2+IJUkdZTaMuvmRarEiIgUmZIYyYvUiqGFzXlIYp6/E+74KlSNLRSmKjGp5dwlI74IejpheEzTahERKRANJ0ledG6rxORhOCnZDrEWqJs95qmF8RhmJViJOfRM2PdUwKKORERkxlASI3nR0dWH2fauujnpbg+64GZQW13Fjo31pVeJaVkSfImISNFoOEnyojPZy7zGeupq8vBPKkO33nSJlgbWdJdYJWbLRvjLp+DF+6OORERkxlASI3nRmeyjNZ+N7jJ0601JNMfoKLVKjBnc/W3tZi0iUkRKYiQvOpN5anTnDid8Cl7xlnEPSbTE6Ez24e65Xy9fYi1QO1vLrEVEikhJjOTM3ens6iWRj+XVZnDYWbDrMeMe0hpvYMvAEN29g7lfL1/MguqRkhgRkaJREiM56+4bZPPAUH72Tep6ER75HfR1j3tIKlnqKLU9lNQrRkSkqJTESM5Sy6vzMpz03B3wmzNg8/pxD0ldp+Q2glTXXhGRotISa8lZakPGvHTrTVUyMuxgnZK6Tmep9Yo58F2w85FRRyEiMmMoiZGcpZKJvFRiulfDrLlQO/65dmysp8oovd2sdz066ghERGYUDSdJzjqTvVQZzG+qz/1kyYl7xADUVFexoDlWenNitmyEh38dzOsREZGCUxIjOevo6mNBc4ya6jw1uhunW2+6RDxWepWYTWvh2rPVK0ZEpEg0nCQ560z2sjBfje52Ox7m7jbpYYmWBh5pH7vLdaRSFaTk6mjjEBGZIZTESM46k33sm2jOz8le98WsDmuNx/jro2txd8xKZNPFWDPUN2uFkohIkWg4SXLi7nQme/PTI2ZgM2x8Boa2TnpoIt5A/+AwL2+Z/NiiUq8YEZGiURIjOenaspW+rcMkWvKwMunFe+GbB8PqyTdRTCVNJbeHkrr2iogUjZIYyUlqhVBeNn/c1iNm4tVJwLakqeR6xez9etjzNVFHISIyI2hOjOQktUIoL5WY7nbAoCkx6aGppKnkuvYedlbUEYiIzBiqxEhOOvNaiVkNjfOhpm7SQ+c11lNbbXSU2jLrrX2w7rFgfo+IiBSUkhjJSUeyj5oqY25jHhrddU/e6C6lqspY0BxjTalVYl68By5/FXQ8FHUkIiIVT8NJkpM1yaDRXXVVHpY5x1ogvlPWh7fGG+gotTkx23rFaHKviEihKYmRnHR09eZn40eAd/x4SocvjMd46MWX83PtfEklMd1qeCciUmgaTpKcdCb78rPxo/uUX5JoibEm2cfw8NRfWzB1s6BhB1ViRESKQEmMTNvwsLMm2UciH5WYdY/CFxfDkzdm/ZLWeANbh5wNm/tzv34+NatXjIhIMSiJkWl7afMAA0PDtOajEpNsh4EeiMWzfkmq4V3JbQS56BCYNS/qKEREKp7mxMi0pZZX52XLgVTlIp7d6iSA1rSGdwdmPx+48E75VtQRiIjMCDlVYszsw2b2iJmtMrNfmlmeZnhKOUh1y83LnJjudrAqaFyY9UsWlmrDO4Dh4WnN8xERkexNO4kxs0XAh4A2d98PqAbela/ApPR1hvsW5WVOTLI9SGCqsy8Ozp1dR11NVeltPfDo9fDfC+Clf0QdiYhIRct1OKkGaDCzrcAsoCP3kCSfVrUnSfYWZqfnB1/ooq6mirmzJ++wO8LAFmhfBj4c3N/pldDTOaWhJAAzIxGPld4mkA0tMDQAT94A3fsFjy0+DOpmw/onoSfDfyZz9wx+/p61sP6xkc/FWqD1oAIHLSJSfqadxLh7u5l9BXgB6AVudPcxS0vM7GzgbIAlS5ZM93IyDS9u3MLJ37qzoNfYe0ETZlNsdHfXZXDbpdvvn78S3vtbGNg05esn4rHSq8TM2R0wuPFT2x/74P2w415wz+XwQIZ+OG/8Ghz2fnjuDrjm/WOf/88HYN4eBQtZRKQcTTuJMbMdgFOBXYEu4Ddm9l53/3n6ce5+BXAFQFtbmyYJFNHzL20B4POnvoK9FzYX5Bo7z5019Rf1rIGGOfCuK4P7s+dDVRXEph5ja7yBe5/dOPUYCim+CP7jbuhNa8QXXxx8P/JcOOCdY18zZ7fg+27Hwfv+vP3xrhdhxa9gqMSWkYuIlIBchpNOAp519/UAZvZb4Ejg5xO+SoqmI5zweuxe81kynWSjUPp7YNYc2PnInE+VaImxpruPoWHPz9YH+TJ/aebH5+4efI1n9rzgK2Vn4MB/zmtoIiKVIpfVSS8ArzKzWRaMJ5wIPDbJa6SIUv1TFsTzsDljPu1+PBz47rycamG8gaFhZ31PBVcqBgegLxl1FCIiJSeXOTH3mtnVwIPAIPAQ4bCRlIbOZC/zGuupr6mOOpSRDjk9b6dqDZdZdyR7ty25rjjfOgR2ORre8t2oIxERKSk5rU5y988Cn81TLJJnHcm+/G3OmE9rHw1W8DS35nyqVI+azq4+qNR5400JbSgpIpKBth2oYJ1dvfnppptvV74Dbv58Xk6VStJKsuFdvsQXaUNJEZEMlMRUsDX52mE63/p7prUSKZN4Qy0NtdWlt8w6n1IbSqoDsIjICEpiKlRP31Z6+gdLrxIzPAz93VCfnyTGzEi0xCq8ErMYBvtgS4ktJRcRiZiSmAq1bV+jlhKrxGzdDDjUN+XtlEHX3gquxMQXB7ti9yqJERFJp12sK1SqFX9rqVVi+nuC73kaToJgcu8dT63P2/lKztI3BV8iIjKCKjEVqmQrMcODsHD/YMVNnrTGY6zr6Wfr0HDezikiIqVPSUyF6uzqpcpgQVOJNbprWQLn3Al7/VPeTploacAd1nZX6JCSO3zvaLjty1FHIiJSUpTEVKjOZB/zm2LUVFf+rzg1eXlNpa5QMoO+LtjwZNSRiIiUlMr/hJuhOpN9pdnB9pHfwf/sBBueytspW8Mhs45KTWJg+zJrERHZRklMhepI9pZmt96+ZLDEujZ/c3VSyVpnVwUvs25eBEl17RURSackpgK5O51dJdzoDvK6xLo5VktjfU1lN7yLL4LujqDPjoiIAEpiKlKydyu9W4dKr9EdBFUYgLr8JTGQ6hVTyZWYxTC8FTZX8FJyEZEpUp+YCpSqSLSW2vJqCCoxdU1Qld/8OdHSUNmVmAPeAUtPhtk7Rh2JiEjJUCWmAqVa8JfkxN7+7rw2uktpjccqO4lp2CHY9TvPyZ+ISDlTJaYCpVrwt5binJiTvxFuPZBfiXgDGzb10z84RH1Ndd7PH7mtvfD782Cfk2HfU6KORkSkJOjPugrUmeylpsrYsdQa3QFU10AsnvfTpub/rE325/3cJaEmFixPX31f1JGIiJQMJTEVqLOrjwXNMaqrLOpQxvrTx+C2L+X9tIlwOXlHpe5mbRYMJ3V3RB2JiEjJ0HBSBepI9pbmyiSAZ26F+fvm/bSp5eSdlZrEQLCbdVIN70REUlSJqUBrkn2lt/FjSn9PXnvEpKQa+6XmA1Ukde0VERlBSUyFcXc6k32lW4np6y7InJhZdTXEG2ord/8k2D6cNDwUdSQiIiVBw0kVZuPmAfoHh0sziRkeClYmFaASA8Hk3ooeTjrgn2HJEcGu1iIioiSm0qR6pZTmlgNht94CJTGtLQ2VPZw0f5/gS0REAA0nVZxU6/2S3Pyxdha89xrY540FOf3CSq/E9HXD378Na1ZGHYmISElQElNhSroSU1MPe5wEO+xSkNO3xmO8vGUrvQMVOmdkeBBu/BQ8e3vUkYiIlAQlMRWmM9lHXXUVc2fXRR3KWC8/D3d8tWDLhCt+mXXDDkE1S8usRUQAJTEVpzPZy4J4PVWl2Ohu/RNw88XQs6Ygp081vKvYFUpm4TLr1VFHIiJSEpTEVJjOrr7SHEqCwk/sDX/ujkpNYgDii9S1V0QkpCSmwnQke2ktxeXVUPAkJrVrd2dXhQ4nATSra6+ISIqWWFeQ4WFnbXeJd+sFiDUX5PSx2mrmzK6r7ErM0pNh7u5RRyEiUhKUxFSQDZv62TrkpVuJ6esGqwompxZIxTe82/v1wZeIiCiJqSQlvbwaYOcjwD4aTFAtkES8gRc3binY+SPXvwnaH4D5S6FxftTRiIhESnNiKkiqArGwVCsxe5wEJ3y6oJdobanwSkzXC/DTU9QrRkQEJTEVJdVyv7VU58SsWQVrHy3oJRLxBrr7BtncP1jQ60Qmvij4rt2sRUSmn8SY2d5mtjztq9vMzs9jbDJFncle6muq2GFWbdShZHbDhfCHDxf0EqntFiq2GhOLQ12TViiJiJBDEuPuT7j7Qe5+EHAosAW4Nl+BydR1JPtobWnACjjnJCf9PQVbmZSysDlIYip6I8j4IlViRETI38TeE4F/uPvzeTofAH1bh/jW357K5ykr2vIXuth5buFW/uSsv6fgy4NTQ2kVW4mBoGtv+wPB7c0vwT3fGXtMwxw48j+D23d8DQY2jT3mqPODpHLVNbD2kbHPv+ItsHD/4FqP/3Hs860Hw9I3lWcMx30Sqku0YikiWctXEvMu4JeZnjCzs4GzAZYsWTKlkw4MDfP9257JObiZ5J1tO0Udwvj6u6G+sJWY+c31AKzv6S/odSK1y6uhJfw9974Md31j7DFzdt/+4X3/D2FThq0eDvu3IIF48i9BEjHa/H2DBGLNyszXOOT0IIEoxxiO+ZiSGJEKYO6e2wnM6oAO4BXuvnaiY9va2nzZsmU5XU/K2BcWwOFnw2s/X9DL7PXpP/O+o3bhk69fWtDryIxUomO1IjNTPlYnvR54cLIERma44SFYfBjM27Pgl2qqr2FTX4WuThIRkW3yMZz0bsYZShLZpqoazvxDUS7VFKuhR0mMiEjFy6kSY2azgdcAv81POFKxhoeDryJojNWwqVL7xIiIyDY5JTHuvtnd57p7Ml8BSYVauxIu3gEe/1PBL9VYX0NP39aCX0dERKKljr1SHH3dwfe62QW/VFOsVsNJIiIzgJIYKY7+nuB7gZvdQTixV8NJIiIVT0mMFEcqiSlwnxjQxF4RkZlCSYwUR384nFSEJCY1sTfXHkgiIlLalMRIcWzdEnyvbyr4pRrraxkadnq3DhX8WiIiEh0lMVIcR50H//US1MYKfqmmWND+SA3vREQqm5IYKZ7qfG3VNbFUEtOjyb0iIhVNSYwUx9/+G377gaJcalsSo0qMiEhFK86fxiJrVkB3R1Eu1Vgf7E6s4SQRkcqmSowUR39PUVYmQdCxF1DXXhGRCqckRoqjr7soK5NAc2JERGYKJTFSHP3dRenWC1qdJCIyUyiJkeLoL14lZvtwkpIYEZFKpom9UhynXQMNLUW5VE11FQ211Wzq15wYEZFKpiRGimPxoUW9nPZPEhGpfBpOksLrS8JNn4WO5UW7ZGOsRhN7RUQqnJIYKbxN6+Guy2DDk0W7ZFN9jSb2iohUOCUxUnhF3ME6pSlWqz4xIiIVTkmMFN62JKY4q5MgWKG0ScNJIiIVTUmMFF5/T/C9iEmMJvbKTGBmbmZfTbt/gZldlKdz/5+ZvT0f55rkOu8ws8fM7JYMz+1lZn8ys6fM7EEzu8rMFhQ6prTrn29ms8Z57lYze8LMHjazu8xs7xyuc6aZfTu8fY6ZnT7BsbuY2XvS7reZ2Tene+0s47vMzI4xs2vNbLmZPW1myfD2cjM7Msvz/D2LY35lZntmG5uSGCm8VBJTpGZ3EEzs1ZwYmQH6gbea2byoA0lnZlNZ+fp+4N/c/fhR54gBfwS+6+57uvshwOXAjtOJYYoxpZwPZExiQqe5+4HAT4AvZ4iheqoXdPfvuftPJzhkF2BbEuPuy9z9Q1O9TrbMbC7wKne/3d3f4u4HAWcBd7j7QeHX38NjJ3yP3T2bZOe7wMezjU9JjBTewv3hhP+CWcX7/2xTfQ2bBgYZHvaiXVMkAoPAFcCHRz8xupJiZpvC78eZ2W1mdp2ZPWNml5jZaWZ2n5mtNLPd005zkpktM7Mnzezk8PXVZvZlM7vfzFaY2QfSznuHmV0PPJohnneH519lZpeGj30GeDXwQzMbnQS8B7jb3X+fesDdb3X3VWYWM7Mfh+d7yMyOD893ppldb2Z/A27OcH+2mf0o/FkfMrNT036mr4SxrTCzc83sQ0ArcEumKtEotwN7pN5nM/uqmT0MHGFm7w2vt9zMvp9KbMzsfeH7eh9wVNr7dJGZXRDe3sPM/hpWex4MfzeXAEeH5/tw+L7/ITx+jpn9LvwZ7jGzA9LO+aOwevRM+LMRvh9/DM+/ysz+OcPP9jbghvF+8AzvcaOZ3RzGuzL1Hqfem/D7cWEsV5vZ42Z2pZlZeNgdBP/usko61SdGCm/h/sFXETXFanGHzQODNMVqi3ptmXl2ufCPlwEH5fm0y5+75I3nZ3Hcd4AVZvalKZz7QGApsBF4BviBux9uZucB5xJUICD4q/9wYHeCD/M9gNOBpLsfZmb1wF1mdmN4/CHAfu7+bPrFzKwVuBQ4FHgZuNHM3uzuF5vZCcAF7r5sVIz7AQ+ME/8HAXf3/c1sn/B8e6XFcIC7bzSzM0fd/yLwN3f/VzNrAe4zs7+GP9MuwEHuPmhmc8LjPwIc7+4bJnk/3wSsDG/PBu5194+a2VLgE8BR7r7VzC4HTjOzm4DPhe9HErgFeCjDea8ELnH3ay2oTFUBF4bvVyqpPC7t+M8BD7n7m8P39ads/3e5D3A80AQ8YWbfBV4HdLj7G8NzxTPEcBRw9SQ/f/p7XAO8xd27LagQ3mNm17v76L8oDwZeAXQAd4XXudPdh83saYJ/o+P9/rdRJUYKb+0j8MxtRb1kY2r/JE3ulQrn7t0EH1ZTGVK439073b0f+AeQSkJWEnyYp1zl7sPu/hRBsrMP8FrgdDNbDtwLzAVScxjuG53AhA4DbnX39e4+SPDhfMwU4h3t1cDPAdz9ceB5IJXE3OTuG9OOTb//WuDCMPZbgRiwBDgJ+H4YG6NeP5Erw3MdBVwQPjYEXBPePpEgUbk/PO5EYDfglWx/PwaAX48+sZk1AYvc/dowpj533zJJPK8GfhYe/zdgrpmlxvH/6O79YUK2DlhA8Pt+jZldamZHu3sywzkTwPpJrpv+HhvwRTNbAfwVWBRea7T73H21uw8Dyxn5724dQRVsUqrESOHd+3144s/wsaeKdsltO1n3DZLI9LeFSB5lWTEppMuAB4Efpz02SPiHqplVAXVpz/Wn3R5Ouz/MyM+F0X89O8GH1Lnu/pf0J8KKwObpBD+OR4Bjp/G60TGk3zfgbe7+RPoB20cypuy0DBWkPncfSrveT9z9k6Ou9+bpXjAH6b/zIaDG3Z80s0OANwBfMLOb3f3iUa/rJUj2JpL+Hp9GMG/p0LD69Nw4rx8TT9r9WHjdSakSI4XX31PUSb2gTSBlZgn/Cr6KYJJsynMEVQCAU4DpjKu+w8yqwrkYuwFPAH8B/t3MamHbCqLZk5znPuBYM5sXzgl5NzBZefYXwJFm9sbUAxaskNmPYN7EaanrE1RTnsh4lpH+Apybmn9hZgeHj98EfCA1D8PM5oSP9xAMv0zXzcDbzWx+6rxmtjNBBetYM5sbvo/vGP1Cd+8BVqcSHjOrt2Cl1EQxpb8vxwEbwkpdRuEw3xZ3/znBxORDMhz2GOF8nyzFgXVhAnM8sPMUXpuyF7AqmwOVxEjh9fcUdXk1sG0ejBreyQzyVSB99vz/EnxQPgwcwfSqJC8QJCB/Bs5x9z7gBwQTdx80s1XA95mkqu/unQRzOW4BHgYecPfrJnlNL3AyQdLxlJk9CvwHwdDG5UCVma0kGIo5Mxwam8znCZK5FWb2SHif8Gd6IXz8Ybav/rkCuMEmn9g73s/wKPBpgjk7KwiSpUT4flwE3E0wH+SxcU7xL8CHwtf+HVgIrACGwsm4oyd0XwQcGh5/CXDGJCHuTzAvaDnwWeALGY75I3DcJOdJdyXQFv5uTgcen8JrsWAJfa+7r8nq+LFzbQqnra3Nly0bXXmTivfD10JNDM64vmiXfHJtD6/9+u18+z0Hc/IBWQ2timRj2uMOIuXKzO4ETnb3riJc68NAt7v/MJvjVYmRwougEpMaTlKvGBGRnH2UYMiuGLoI+u5kRRN7pfAWHwZzdi3qJdMn9oqIyPS5+71FvNaPJz9qOyUxUninFLQjdkaz62owgx4tsRYRqVgaTpLCcofeLhgemvTQfKqqMhrrajSxV0SkgimJkcIa2ASX7gx3f6fol9b+SSIilS2nJMbMWtL2PnjMzI7IV2BSISLYwTqlsb5GHXtFRCpYrpWYbwA3uPs+BPscjLfWXWaqCJOYpliNJvZKRTMzN7Ovpt2/wMwuytO5R2wgWShm9o7wj+AxvVjCRnp/CvvEPGhmV4V9RIrCzM4PG8xleu5WM3si7Ndyl5ntncN1zjSzb4e3zzGz0yc4dhcze0/a/TYzK+jEQzO7LGw0eK0FG08+bWbJ8PZyM8tmd+pMse9vZv+XS2zTntgbbhR1DHAmQLj/w0AuwUiZWXk13PPdsY+/7n9gp8Nh+S/h7+F/W7Hi9/5vjNWS7M1+Tsxnr1vFw6szbR0ys9RVV/GFt+zHXguKn3jKlPUDbzWz/8lik8KiMbOa1D5EWXg/8G/ufueoc8QIGq19JLWTddiFdkdg7VRjmGJMKecT7NE03p5Fp7n7MjM7m6Dj7SmjYqhO24IgK+7+vUkO2YWgGd8vwuOXAQVrwGZmc4FXufv5BLt1p34P2zahnIJdGBn7SjNbbGZL3P2F6cSXy+qkXQk6J/7YzFK7TZ7n7iO6Qoa/3LMBliwp1jJzKYp1j0L7Mtj9xJGPV4X/rGrqoCkBc/eARYeOfX2BNcVqWP3yZPulBQaHhvnZPc+z89zZ7DQn4x9eM8LwsHPn0xu4/cn1SmLKwyBBV9kPA59KfyL8C/cP7n51eH+TuzeGH0CfI+jHsT/BdgUrgfOABuDN7v6P8DQnmdmFQDNBMvGHcNuASwi6uNYD33H374fn/TzBLtX7sH1DxlQ87wb+H0HDwD+6+yfM7DMEmxb+0IKdjj+W9pL3AHenEhgAd781PFcM+C7QFr4HH3H3WyzYtfqtQCNQbWY/HnX/DcC3CHbIrgUucvfrwp/pUoJdnYcJuh0bwSaEt5jZBnc/foLfw+2EO3+b2SaCLsYnAR80s10INuesI9hu4D/cfcjM3gd8Mvw9PEy4l1BYSdvk7l+xYNfw7xEkbkME2xNcAiwNu+z+hGD36wvc/eRwu4QfEWwRsQU4291XhOdcEj6+BLjM3b9pwXYRVwGLgWrg8+4+ejPKtwE3jPeDm9mOYYypD/jz3f0uMzuWYLQGgj23jhkdu7t/Hfg98C5gKruwb5NLElNDsM/Cue5+r5l9g6Ct9H+lH+TuVxD8R0ZbW1vx2gNL4Z34meBrPPu9LfiKSFN99hN71/X0M+xw9jG78e7DZ26y7e4s/cwNrEn2RR1K+bkofmvmx5PHhc9fBhyU4YjzuSi5nIviZxJWtke8bnLfIWiXP5UPgQOBpcBGgt2pf+Duh5vZecC5hB/IBH85Hw7sTvBhvgdBK/mkux9mZvXAXWaW2gX7EGC/0TtZh3v0XEqwl9PLBG343+zuF5vZCQQfwqOrCfsR/HGcyQcBd/f9zWyf8HyppOkQ4AB33xgmNen3vwj8zd3/1cxaCFru/zX8mXYBDnL3QTObEx7/EeD4LKpcbyJIBAFmA/e6+0fNbCnwCeCocC+hy4HTzOwmgkTyUCBJsB3DQxnOeyVwibtfGyZuVQSfs9uqIGHymPI54CF3f3P4vv6U7f/m9gGOJ9h36Qkz+y5B0tbh7m8Mz5WpZH4UcPUEP/s3gK+7+51mtoRgf6qlBLt6fzBMaBqBvtGxh5aFjxc9iVkNrE5rgnN1GIjMFEODUF26rYYa67OfE9OZDDZMXRifbLPWymZmJOINdCqJKRvu3m1mPyX4az+rnX+B+8P9ezCzfwCpJGQlwQddylXuPgw8ZWbPEHwQvhY4IG2+TBzYk2A6wX2jE5jQYcCt7r4+vOaVBH+Z/y7LeEd7NUFFBXd/3MyeZ3vl56ZwQ0wy3H8tcIqZXRDejxFUEE4Cvpcabhr1+olcaWa9BJttnhs+NgRcE94+kSBRud+CPScbgHXAKxn5fvyasZWrJmCRu18bxtQXPj5RPK8mqJzg7n+zYIPJ1O67fwz3l+o3s3XAAoLf91fN7FKCqt0dGc6ZIBh1Gc9JwL5pcTWHSctdwNfC3/Vv3X31OLGvI6h4Tcu0P4HcfY2ZvWhme4fbmp9IsCmYzAT9m+CSneB1l8ArPxB1NBk1xWrp3TrE4NAwNdUTz2Hv6Ao+tFvjDcUIraQl4jE6ktl+Fso2k1VOLkqeP8nz/wf83zSvfhnwIJDe7XSQcPGGmVURDGekpG+WOJx2f5iRnwujq+dOMMxyrrv/Jf2JsCIwnU0mx/MIcOw0Xjc6hvT7Brwt/Mza/uDEicFETstQQepLmwdjBMMmnxx1vTdP94I5SP+dDwE17v6kmR0CvAH4gpnd7O4Xj3pdL0GyN54qgjkzo//yucTM/hie+y4z+6dxXh8j++Q748VzcS5BJrqCoGT1xRzPJ+Wiux18GBrmTH5sRBrDrQc2908+ry5ViUm0zOxKDBBUYrpUiSknYeXgKoJJsinPEVQBIJhwWjuNU7/DzKrMbHeC+RRPEAwX/LuZ1cK2FUSzJznPfQQ7as8L55+8G7htktf8AjjSzN6YeiBcIbMfcAdwWur6BNWUJzKeZaS/EOyKbeFrDw4fvwn4gJnVhI+n/sfWQzD8Ml03A283s/mp85rZzgRzY44NKyW1BHNdRnD3HmB1KuExs3oLVkpNFFP6+3IcsMHdu8cLLhzm2+LuPyeYmHxIhsMeA/aY4Ge8ke1VKMzsoPD77u6+0t0vBe4nqOJlin0vYNUE559QTkmMuy939zZ3P8Dd3+zuL+dyPikjydXB9/iiaOOYQGr/pO4suvZ2dPXRWF9Dc2w6/5+vLK0tMdb19DE4NBx1KDI1XwXmpd3/X4IPyoeBI5heleQFggTkz8A54V/bPyCouj9oZqsIJrFOWNUPh64uJJj78TDwgLtfN8lreoGTCZKOp8zsUeA/CIY2LgeqzGwl8GvgzHCoZDKfJ0jmVpjZI+F9wp/phfDxhwkmFUMwn/MGy7D8Oxvu/ijwaYI5OysIkqVE+H5cBNxNMOwyXnuSfwE+FL7278BCYAUwZMHS7g+POv4i4NDw+EuAMyYJcX+CeUHLgc8CX8hwzB8JJnGP50NAm5mtCH9H54SPn29mq8JYthL8G8oU+/HhNabF3Is317atrc2XLSvYSjAppgd/CtefC+etgB12jjqajP68spN/v/JB/nze0SxNNE947Ad+toxn1m/mpo9Mp3pdWX5x7wv8v2tXcteFJ7CoRcNro0x73EGkXJnZncDJ7t6V5/PWE1TkXj2N5e+Ath2Q6eruAAyapz0fq+CawqpKNpN71yT7SOgDG9g+pLZG82JEJPBRti+hzqclwIXTTWBASYxM15aN0LgAqkt3+CU1J2ZTfxbDSck+Es2aDwPbJzd3aF6MiADufq+7ryjAeZ9K9f6ZLiUxMj1v+BKcn/d/03nVWB8kMZNVYgYGh9mwqV+TekOpZeadqsSISIlTEiPTV1MfdQQTao5ll8Ss7e7DXcurU5pjNcyuq1YlRkRKnpIYmTp3+MFJsOzHkx8boe3DSRMnMR1dWl6dzsxItDSoEiMiJU9JjExdXxJW3799h+oS1VBbTXWV0TPJEutUd9qEKjHbJOIxde0VkZKnJEamrrs9+F7CPWIgqCg0ZrF/0vYkRpWYlFZtPSAiZUBJjExdMkximhdHG0cWGutr6JlkOKkz2RvMA6kv3X2gii3REmPDpn4GBtXwTkRKl5IYmbru0u/Wm9IUm3wTyI6uPlrVI2aE1ngD7sGkZxGRUqUkRqYu2Q5WBY0Lo45kUk2xbIaTejWUNEpqmXVq0rOISClSEiNT96r/gLNuhurSH35pitXSM0mzu0516x2jtSXVK0aVGBEpXaX/KSSlZ/bc4KsMNNbX8Mz68SsxfVuH2Lh5gFZVYkZIrdRSEiMipUyVGJm6G/8LVl0TdRRZaYrVTNgnZk34Ib1Qy6tHmF1fQ3OsRr1iRKSkKYmRqXGH+/4X2h+MOpKsNMZq6J5gTkxH+CGtSsxYrS0N6torIiVNSYxMTe/LMNgLzaW/Mgmgqb6GgcFh+geHMj7fGX5Ia07MWEHDO1ViRKR0KYmRqUmWz/JqCCb2AuOuUEp9SGt10lgL1fBOREqckhiZmu7yaXQH23eyHm9eTEeyjzmz64jVVhczrLLQGo+xcfMAfVszV7FERKKmJEamJlWJaW6NNo4sNU2yk/WaZJ+qMONIDbGtUTVGREqUkhiZml2PgTd8BRrnRx1JVhonSWI6utTobjypyc4dmhcjIiVKfWJkanbcO/gqE0314ZyYcYaTOpN9HLbLnGKGVDZSlZhOrVASkRKlSoxMzSO/g+fujDqKrG0fThrbtXfLwCDJ3q0kWlSJySRVodIKJREpVUpiZGr++llY9uOoo8haajgpUyUm1QOlVY3uMorVVjNndh0dmhMjIiVKSYxkb3gYujvKZnk1TDyxV8urJ7ewOUanNoEUkRKlJEayt+UlGBoom+XVAPU11dRVV42TxISVGDW6G1drS0y9YkSkZCmJkex1l1eju5TGWA2bMuxknZqwOr+5vtghlY2EGt6JSAlTEiPZS6Ya3ZVHj5iUpljNuMNJ8xrrqa9Ro7vxJFpiJHu3smVg/P2nRESioiRGsteUgENOh5ado45kShrrazJuO9CR7KNVK5MmlJr0rI0gRaQUqU+MZG/xocFXmRm3EtPVy247zo4govKRvsx6j/mNEUcjIjKSKjGSvfVPQNcLUUcxZY31tfRkWGIdbDmgSb0TSU161rwYESlFSmIke9d/CK7996ijmLLmDBN7e/q20tM/qOGkSaQmPatrr4iUIiUxkr3u9rJbmQTB6qTRw0mpysJCVWImVF9TzbzGenXtFZGSpCRGsjM8FDS6ay7DJCac2Ovu2x7rCBu4tarR3aRaW2Lq2isiJSmnib1m9hzQAwwBg+7elo+gpARtWgs+VJaVmKZYLYPDTt/WYRrqguXUqUpMQo3uJpWIx3hm/eaowxARGSMfq5OOd/cNeTiPlJKhwaA7L0BtQ1CFgbLq1puS2j9pfU8/OzYFczxe3LiFKoMFTWp0N5lEvIG7nn6J3oGhEY/Haqsws4iiEhHREmvJZLAfLjsANq0J7l/wNDTsALufAC07RRvbNLQ01AJwzJdvGfF4azxGTbVGVCezeIcGNvUPsvQzN4x4/OQDEnz7PYdEFJWISO5JjAM3mpkD33f3K0YfYGZnA2cDLFmyJMfLSVEkVwcJzH5vh4X7Q90saNwRjvk4zN836uim7IR95vPZN+1L/+DwiMcPWByPKKLy8rZDgurb4PD2OUV/WtnJ8he7IopIRCSQaxLzandvN7P5wE1m9ri7355+QJjYXAHQ1tbmmU4iJaY73F7gkNNht2O3P77zEdHEk6PZ9TW876hdow6jbO0wu46zjt5txGNdW7bywzufYXjYqarSkJKIRCOnWrq7t4ff1wHXAofnIyiJWGqPpHj5zX+R4mhtibF1yNmwqT/qUERkBpt2EmNms82sKXUbeC2wKl+BSYR2PQbe8RMlMTKuVKdjLb0WkSjlMpy0ALg2XJ1QA/zC3W+Y+CVSFuKLynIptRRPak+lNcle2Kkl2mBEZMaadhLj7s8AB+YxFikVy38ZTObd99SoI5ESlUpitLu1iERJ60tlrLu/HSQyIuOYM7uO+poqbUcgIpFSEiNjJVdrOEkmZGYk4tqOQESipSRGRhrYDH1dZblHkhRXIt5AZ5cqMSISHSUxMlJqewGtTJJJJFpi2/agEhGJgpIYGSm5OviuSoxMIhGPsa6nn8Gh4ckPFhEpACUxMlJ8MRz9UZi7R9SRSIlLxBsYGnbWq+GdiERESYyMNG9POPEz0LQg6kikxLW2aJm1iERLSYyM1PEQdCyPOgopA6muvVpmLSJRURIjI938efj9eVFHIWWgNZXEqBIjIhFREiMjdbdrZZJkpbmhhll11XSoEiMiEVESIyN1d2hlkmTFzFgYj7FGy6xFJCJKYmS7vm7o71a3Xslaa7xBXXtFJDJKYmS77vbguyoxkqVEPKauvSISGSUxksZgj9cEy6xFspBoaWD9pn4GBtXwTkSKrybqAKSEzN8H3nt11FFIGWmNx3CHtd197DRnVtThiMgMo0qMbNf7MmzV0IBkL9ESLLNe0615MSJSfEpiZLu/fAq+eXDUUUgZScRTXXuV/IpI8SmJke262zWpV6YklcRoN2sRiYKSGNku2Q7NrVFHIWWkKVZLU32NViiJSCSUxEjAXd16ZVoSLTH1ihGRSCiJkUDvy7B1i4aTZMoS8QZtAikikVASI4Hel6FxAbTsFHUkUmZaW7T1gIhEQ31iJDB3d7jgyaijkDK0sLmBDZsG6B8cor6mOupwRGQGUSVGRHKSaAlWKKkaIyLFpiRGArd/GS4/MpjgKzIFrfGg4V1Hl5IYESkuJTES2PB0sIO1WdSRSJlJVWI0uVdEik1JjATU6E6mSQ3vRCQqSmIk0N0OcSUxMnWz6mqIN9SqEiMiRackRsJGdx3q1ivTlojH6NScGBEpMiUxAltegsE+aFa3Xpme1pYGde0VkaJTnxiBWXPhgqehujbqSKRMJeIxHnzh5ajDEJEZRkmMBCuSGneMOgopY60tDXRt2UrvwBANdWp4JyLFkfNwkplVm9lDZvaHfAQkEXjs93D1+2Fgc9SRSJlKrVDq0OReESmifMyJOQ94LA/nkaisXgaPXgc1DVFHImVqYVxde0Wk+HIaTjKzxcAbgf8GPpKXiKR4nrkVtvZCx4PByqQqzfOW6Ul17b3l8XX0DgzlfL6GumqO2G0uVVVqvigi48t1TsxlwMeBpvEOMLOzgbMBlixZkuPlJGc9a2B4KEhafn8evPxc8Pger4k0LClvC+MxZtVV84M7n+UHdz6bl3NeedYrOWqPeXk5l4hUpmknMWZ2MrDO3R8ws+PGO87drwCuAGhra9PGPFG746vw8K/gky/Cu34BQwPB43N2jzYuKWux2mpu/uixbOgZyPlcG7cMcMaP7uO5lzYriRGRCeVSiTkKOMXM3gDEgGYz+7m7vzc/oUlBJNO2F1jwimhjkYqSiDeQiOc+r2po2KmuMs2vEZFJTXsShLt/0t0Xu/suwLuAvymBKQPaXkBKXHWVMb+pXrtii8ikNJNzptFGj1IGEvGY9mISkUnlJYlx91vd/eR8nEsKaLAfNq9XEiMlL9HSoF2xRWRSqsTMJH1JWLAfzNUkXiltrfEYHV29uGstgIiMT9sOzCSN8+Hf74o6CpFJJeIN9A8O8/KWrcyZXRd1OCJSolSJmUn0V62UidaWcBuDLs2LEZHxKYmZSe78Oly6C2zVXAMpbQvDpdpaZi0iE1ESM5MkVwfVmNpY1JGITKg13ItJK5REZCJKYmaS7naIL446CpFJzWusp7ba6FAlRkQmoCRmJkmqR4yUh6oqY0FzjE7NiRGRCSiJmUnUrVfKSGu8QZUYEZmQkpiZYmsf9HWpEiNlY6G69orIJNQnZqaojcGn18HQ1qgjEclKoiXGn1f1MTzsVFVZ1OGISAlSJWYmqa6FullRRyGSldZ4A1uHnJc2D0QdioiUKCUxM8Xjf4SfnAKbN0QdiUhWElpmLSKTUBIzU6x9FJ69Deoao45EJCutLUHDu44uTe4VkcyUxMwU3ath1jw1upOyoUqMiExGScxMkdTyaikvc2bXUVdTRaeWWYvIOJTEzBTd7dCsbr1SPsyMRDymTSBFZFxKYmYKNbqTMpSIx7QJpIiMS31iZgJ3OP06qG+OOhKRKWmNN3DvsxujDkNESpSSmJnADFoPjjoKkSlLtMRY093H0LBTrYZ3IjKKhpNmgjUr4cZPQ3dn1JGITEki3sDQsLO+pz/qUESkBCmJmQk6HoK/fwuG1PlUyktqmXWHllmLSAZKYmaCZDtg0JSIOhKRKUnEg4Z3nWp4JyIZKImZCbpXQ+N8qKmLOhKRKWltUcM7ERmfkpiZINkOzVpeLeUn3lBLQ221Gt6JSEZKYmYC9YiRMmVmJFpiqsSISEZaYj0THPMxaJgTdRQi09Iab9AmkCKSkZKYmeCAd0Ydgci0LYzHuOOp9VGHISIlSMNJlW7TeljxG9i8IepIRKalNR5jXU8/W4eGow5FREqMkphK1/EQ/PYs2PhM1JGITEuipQF3WNutISURGUlJTKXrbg++a3WSlKlUwzutUBKR0ZTEVLrudrAqaFwQdSQi09LaEja8UxIjIqMoial0yfagU2+15nBLeVqYqsR0aZm1iIykJKbSda/WUJKUteZYLY31NarEiMgY0/7z3MxiwO1AfXieq939s/kKTPJk12OgWtsNSHlLxGN0qBIjIqPkMsbQD5zg7pvMrBa408z+7O735Ck2yYdjPhZ1BCI5S7Q0qBIjImNMO4lxdwc2hXdrwy/PR1AjrH0EHvr52MfnL4VDToetvXDzxWOfr6mHky4Kbt/1TejpHHvMEf8ZtON/7A/w/F1jn9/7DbDr0eUbw2B/UInZ65+gtmHsa0XKRGs8xgPPbeTi3z+an/O1xDjr6N3yci4RiU5Osz3NrBp4ANgD+I6735vhmLOBswGWLFky9YskV2f+8N7zNcGH99BA5ufrm7YnEE/8KUgCRjvoPUEC0b4s8znm7BYkEOUagw/Do9dBfCdYfOjY14qUiSN2n8ufVnbym2Uv5uV8+y2KK4kRqQAWFFRyPIlZC3AtcK67rxrvuLa2Nl+2bFnO1xMRiYhFHYCIbJeX1Unu3gXcArwuH+cTERERmcy0kxgz2zGswGBmDcBrgMfzFJeIiIjIhHKZE5MAfhLOi6kCrnL3P+QnLBEREZGJ5bI6aQVwcB5jEREREcmaOvaKiIhIWVISIyIiImVJSYyIiIiUJSUxIiIiUpaUxIiIiEhZUhIjIiIiZUlJjIiIiJQlJTEiIiJSlpTEiIiISFnKyy7WWV/MbD3w/DReOg/YkOdwikWxR0OxR6PSY9/g7troVqREFDWJmS4zW+bubVHHMR2KPRqKPRqKXUSKScNJIiIiUpaUxIiIiEhZKpck5oqoA8iBYo+GYo+GYheRoimLOTEiIiIio5VLJUZERERkBCUxIiIiUpZKOokxs9eZ2RNm9rSZXRh1PFNhZs+Z2UozW25my6KOZzJm9iMzW2dmq9Iem2NmN5nZU+H3HaKMMZNx4r7IzNrD9365mb0hyhjHY2Y7mdktZvaomT1iZueFj5fD+z5e7CX/3ptZzMzuM7OHw9g/Fz6+q5ndG/7/5tdmVhd1rCIysZKdE2Nm1cCTwGuA1cD9wLvd/dFIA8uSmT0HtLl7WTT+MrNjgE3AT919v/CxLwEb3f2SMIncwd0/EWWco40T90XAJnf/SpSxTcbMEkDC3R80sybgAeDNwJmU/vs+XuzvpMTfezMzYLa7bzKzWuBO4DzgI8Bv3f1XZvY94GF3/26UsYrIxEq5EnM48LS7P+PuA8CvgFMjjqliufvtwMZRD58K/CS8/ROCD6mSMk7cZcHdO939wfB2D/AYsIjyeN/Hi73keWBTeLc2/HLgBODq8PGSfN9FZKRSTmIWAS+m3V9NmfxPMuTAjWb2gJmdHXUw07TA3TvD22uABVEGM0X/aWYrwuGmkhuOGc3MdgEOBu6lzN73UbFDGbz3ZlZtZsuBdcBNwD+ALncfDA8pt//fiMxIpZzElLtXu/shwOuBD4bDHmXLg3HH0hx7HOu7wO7AQUAn8NVIo5mEmTUC1wDnu3t3+nOl/r5niL0s3nt3H3L3g4DFBFXffaKNSESmo5STmHZgp7T7i8PHyoK7t4ff1wHXEvyPstysDec+pOZArIs4nqy4+9rwQ2oY+F9K+L0P52RcA1zp7r8NHy6L9z1T7OX03gO4exdwC3AE0GJmNeFTZfX/G5GZqpSTmPuBPcMVA3XAu4DrI44pK2Y2O5zsiJnNBl4LrJr4VSXpeuCM8PYZwHURxpK1VAIQegsl+t6HE0x/CDzm7l9Le6rk3/fxYi+H997MdjSzlvB2A8HigccIkpm3h4eV5PsuIiOV7OokgHB55mVANfAjd//vaCPKjpntRlB9AagBflHqsZvZL4HjgHnAWuCzwO+Aq4AlwPPAO929pCbRjhP3cQTDGQ48B3wgbY5JyTCzVwN3ACuB4fDh/0cwt6TU3/fxYn83Jf7em9kBBBN3qwn+kLvK3S8O/7v9FTAHeAh4r7v3RxepiEympJMYERERkfGU8nCSiIiIyLiUxIiIiEhZUhIjIiIiZUlJjIiIiJQlJTEiIiJSlpTESNbMbCjcmXiVmf3GzGaVQEzHmdmROZ7jGTPbe9Rjl5nZuJsuhruUz8vluiIikhslMTIVve5+ULhb9ABwTjYvSuuCWgjHAVNKYjLE8yuCZoqp56sImp79KtfgRESkcJTEyHTdAexhZm8ys3vN7CEz+6uZLQAws4vM7GdmdhfwMzPbxczuMLMHw68jw+OOM7PbzOy6sCJyiZmdZmb3mdlKM9s9PG5HM7vGzO4Pv44KNx48B/hwWCE6OtNxmeIZ9bP8EvjntPvHAM+7+/Nm9rtwE89HMm3kGf5cq9LuX2BmF4W3dzezG8LX32Fm2p9HRCSPCvkXslSosJLxeuAG4E7gVe7uZnYW8HHgo+Gh+xJshNkbDj29xt37zGxPgsShLTzuQGApsBF4BviBux9uZucB5wLnA98Avu7ud5rZEuAv7r7UzL4HbHL3r4Sx/WL0ceG5R8ST/vO4+0ozGzazA939YYKqzC/Dp//V3TeG7envN7Nr3P2lLN+qK4Bz3P0pM3slcDlwQpavFRGRSSiJkaloMLPl4e07CPbO2Rv4dbhnTh3wbNrx16clDLXAt83sIGAI2CvtuPtTrenN7B/AjeHjK4Hjw9snAfsGW/YA0BzuoDzaRMddPzqBSfNL4F1m9gjwZoLtCwA+ZGZvCW/vBOwJTJrEhNc8EvhNWiz1k71ORESypyRGpqLX3Q9Kf8DMvgV8zd2vN7PjgIvSnt6cdvvDBHsbHUgwjNmX9lz6/jTDafeH2f5vtIqg4pP+OtISBLI4bvPog9P8iiB5ug1Y4e5rw5/nJOAId99iZrcCsVGvG2TksGzq+Sqga/T7JSIi+aM5MZKrONAe3j5jkuM63X0Y+BeCzfem4kaCoSUAwooOQA/QlMVxE3L3fwAbgEvYPpQUB14OE5h9gFdleOlaYL6ZzTWzeuDk8HzdwLNm9o4wDjOzA7OJRUREsqMkRnJ1EcGQyQMEScB4LgfOMLOHgX2YuCqSyYeANjNbYWaPsn1l1O+Bt6Qm9k5wXDZ+Gcb22/D+DUCNmT1GkNzcM/oF7r4VuBi4D7gJeDzt6dOA94c/8yPAqVOIRUREJqFdrEVERKQsqRIjIiIiZUlJjIiIiJQlJTEiIiJSlpTEiIiISFlSEiMiIiJlSUmMiIiIlCUlMSIiIlKW/j8FyMM77TzLNAAAAABJRU5ErkJggg==)
Markdown is supported
0% or .
You are about to add 0 people to the discussion. Proceed with caution.
Finish editing this message first!
Please register or to comment