2021-06-24 11:07:41 +02:00
{
"cells": [
{
"cell_type": "markdown",
"metadata": {},
"source": [
2021-06-26 17:59:05 +02:00
"# Algorytm najszybszego spadku dla regresji wielomianowej. "
2021-06-24 11:07:41 +02:00
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
2021-06-26 17:59:05 +02:00
"Skład grupy:\n",
2021-06-27 14:09:23 +02:00
"- Nowak Anna,\n",
2021-06-26 17:59:05 +02:00
"- Łaźna Patrycja,\n",
"- Bregier Damian"
2021-06-24 11:07:41 +02:00
]
},
2021-06-26 17:59:05 +02:00
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## 0. Podstawowe informacje o zbiorze danych"
]
},
{
"attachments": {
"image.png": {
"image/png": "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
}
},
"cell_type": "markdown",
"metadata": {},
"source": [
"Ze względu na specyfikę wytycznych projektu zakładającego wykorzystanie jedynie dwóch cech: x i y oraz modelowanie zależności y od x za pomocą funkcji wielomianowej dobrany został przez nas odpowiedni dataset. Obejmuje on jedynie trzy kolumny, czyli: płeć, wzrost w calach oraz wagę w funtach, z czego ze względu na specyfikę projektu wykorzystywane są jedynie wzrost oraz waga. Każdy z parametrów zawiera po 10 tysięcy unikalnych wartości.\n",
"\n",
2021-06-27 17:34:20 +02:00
"\n",
"\n",
"Dokładne informacje na temat tego zbioru, jego zawartości i samych danych można znaleźć pod adresem: https://www.kaggle.com/mustafaali96/weight-height."
2021-06-26 17:59:05 +02:00
]
},
2021-06-27 18:45:49 +02:00
{
"cell_type": "code",
2021-06-28 10:01:22 +02:00
"execution_count": 37,
2021-06-27 18:45:49 +02:00
"metadata": {},
"outputs": [],
"source": [
"import pandas as pd\n",
"import random\n",
"import matplotlib.pyplot as plt\n",
"import numpy as np\n",
"import ipywidgets as widgets\n",
"np.set_printoptions(suppress=True)"
]
},
2021-06-26 17:59:05 +02:00
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## 1. Wczytywanie i preprocessing danych"
]
},
2021-06-24 13:03:10 +02:00
{
"cell_type": "code",
2021-06-28 10:01:22 +02:00
"execution_count": 38,
2021-06-24 11:07:41 +02:00
"metadata": {},
2021-06-25 01:36:15 +02:00
"outputs": [],
2021-06-24 11:07:41 +02:00
"source": [
2021-06-27 11:07:37 +02:00
"# Stopień wielomianu\n",
"degree = 4\n",
"X_plot = np.linspace(0, 100, 1000)\n",
2021-06-25 01:36:15 +02:00
"initial_theta = np.matrix([0] * (degree + 1)).reshape(degree + 1, 1)"
2021-06-24 11:07:41 +02:00
]
},
{
"cell_type": "code",
2021-06-28 10:01:22 +02:00
"execution_count": 39,
2021-06-24 11:07:41 +02:00
"metadata": {},
2021-06-25 01:36:15 +02:00
"outputs": [],
2021-06-24 11:07:41 +02:00
"source": [
2021-06-26 17:59:05 +02:00
"# Wybór dwóch kolumn - dotyczących wzrostu i wagi\n",
2021-06-25 01:36:15 +02:00
"data = pd.read_csv('weight-height.csv')[[\"Height\", \"Weight\"]]\n",
2021-06-26 17:59:05 +02:00
"# Czyszczenie tabeli i wartości pustych\n",
2021-06-25 01:36:15 +02:00
"data = data.dropna()\n",
"data_matrix = np.matrix(data)\n",
"\n",
"m, n_plus_1 = data_matrix.shape\n",
"n = n_plus_1 - 1\n",
"X = (np.ones((m, 1)))\n",
"\n",
"for i in range(1, degree + 1):\n",
" Xn = np.power(data_matrix[:, 0:n], i)\n",
" Xn /= np.amax(Xn, axis=0)\n",
" X = np.concatenate((X, Xn), axis=1)\n",
"\n",
"X = np.matrix(X).reshape(m, degree * n + 1)\n",
"Y = np.matrix(data_matrix[:, -1])"
2021-06-24 11:07:41 +02:00
]
2021-06-24 13:03:10 +02:00
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
2021-06-27 17:34:20 +02:00
"## 2. Metody regresji wielomianowej"
2021-06-24 13:03:10 +02:00
]
},
{
"cell_type": "code",
2021-06-28 10:01:22 +02:00
"execution_count": 40,
2021-06-27 17:34:20 +02:00
"metadata": {},
"outputs": [],
2021-06-24 13:03:10 +02:00
"source": [
2021-06-27 14:09:23 +02:00
"def print_theta(theta):\n",
" print(\"f(x) = \", end=\"\")\n",
" for i,x in enumerate(theta.tolist()[:-1]):\n",
" x = x[0]\n",
" print(f\"{x}x^{i}\", end=\" + \")\n",
" print(f\"{theta.tolist()[-1][0]}^{len(theta) - 1}\")\n",
"\n",
2021-06-26 17:59:05 +02:00
"# Implementacja wzrosu na regresję wielomianową\n",
2021-06-24 13:03:10 +02:00
"def polynomial_regression(theta, x):\n",
2021-06-27 11:07:37 +02:00
" x = x/data[\"Height\"].max()\n",
2021-06-25 01:36:15 +02:00
" return sum(theta * np.power(x, i) for i, theta in enumerate(theta.tolist()))\n",
2021-06-24 13:03:10 +02:00
"\n",
2021-06-26 17:59:05 +02:00
"# Implementacja wzoru na RMSE, czyli pierwiastek z błędu średniokwadratowego\n",
2021-06-25 01:36:15 +02:00
"def mean_squared_error(theta, X, Y):\n",
" J = 1.0 / (2.0 * m) * ((X * theta - Y).T * (X * theta - Y))\n",
2021-06-27 17:34:20 +02:00
" return J.item()"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## 3. Wybrane rodzaje gradientów"
]
},
{
"attachments": {
"image.png": {
"image/png": "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
}
},
"cell_type": "markdown",
"metadata": {},
"source": [
"Metoda gradientu prostego to algorytm numeryczny mający na celu znalezienie minimum lokalnego dla zadanej funkcji celu. Jak sama nazwa wskazuje jest to jedna z prostszych metod, która opiera się na iteracyjnym poszukiwaniu minimum. Proces ten (dla pierwszych czterech kroków), został pokazany na poniższym rysunku.\n",
""
]
},
2021-06-27 20:23:26 +02:00
{
"attachments": {
"image.png": {
"image/png": "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
}
},
"cell_type": "markdown",
"metadata": {},
"source": [
"Metoda gradientu prostego\n",
""
]
},
2021-06-27 17:34:20 +02:00
{
"cell_type": "code",
2021-06-28 10:01:22 +02:00
"execution_count": 41,
2021-06-27 17:34:20 +02:00
"metadata": {},
"outputs": [],
"source": [
2021-06-26 17:59:05 +02:00
"# Wzór na gradient prosty\n",
2021-06-27 11:07:37 +02:00
"def gradient(theta, X, Y):\n",
2021-06-27 17:34:20 +02:00
" return 1.0 / len(Y) * (X.T * (X * theta - Y)) "
]
},
2021-06-27 20:23:26 +02:00
{
"attachments": {
"image.png": {
"image/png": "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
}
},
"cell_type": "markdown",
"metadata": {},
"source": [
"Metoda najszybszego spadku jest iteracyjnym algorytmem wyszukiwania minimym zadanej funkcji celu i stanowi modyfikację gradientu prostego. Na samym początku algorytmu wybierany jest punkt startowy, w którym jest obliczany antygradient funkcji celu stanowiący kierunek poszukiwań w procedurze. Kolejny krok jest obliczany w oparciu o poprzedni i jeżeli nie spełnia on warunku stopu to obliczenia są powtarzane do skutku. \n",
"\n",
"Podstawową różnicę w stosunku do gradientu prostego stanowi sposób wyszukiwania wartości, która jest dokonywana przez minimalizację kierunkową funkcji. \n",
"\n",
""
]
},
{
"cell_type": "code",
2021-06-28 10:01:22 +02:00
"execution_count": 42,
2021-06-27 20:23:26 +02:00
"metadata": {},
"outputs": [],
"source": [
"# Steepest descent\n",
"def steepest_descent(X, Y, theta, cost_function = mean_squared_error, \n",
" single_step=0.01, epochs=30):\n",
" cost = cost_function(theta, X, Y)\n",
" logs = [[cost, theta]]\n",
" for i in range(epochs):\n",
" direction = gradient(theta, X, Y)\n",
" j = 0\n",
" while(True):\n",
" j+=1\n",
" next_theta = theta - single_step * direction\n",
" next_cost = cost_function(next_theta, X, Y)\n",
" if(next_cost >= cost or type(next_cost) != float):\n",
" break\n",
" cost = next_cost\n",
" theta = next_theta\n",
" logs.append([next_cost, theta])\n",
" return theta, logs"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"Ponadto w toku prac projektowych przetestowane zostaly także inne algorytmy takie jak batch gradient descend (BGD), BGD z momentum, mini-batch gradient descend oraz stochastic gradient descend. \n",
"\n",
"BGD to metoda w której błąd obliczany jest dla każdego parametru ze zbioru treningowego, ale sam model jest aktualizowany dopiero po ocenie wszystkich danych. \n",
"\n",
"MBGD opiera się na podziale danch treningowych na niewielkie pakiety, które są potem wykorzystywane do obliczania wartości błędu i aktualizacji modelu.\n",
"\n"
]
},
2021-06-27 17:34:20 +02:00
{
"cell_type": "code",
2021-06-28 10:01:22 +02:00
"execution_count": 53,
2021-06-27 17:34:20 +02:00
"metadata": {
"scrolled": true
},
"outputs": [],
"source": [
2021-06-27 20:23:26 +02:00
"# Batch gradient descend (BGD)\n",
"def BGD(X, Y, theta, cost_function = mean_squared_error, learning_rate=0.1, \n",
" eps=10**-5, max_steps = 10000000000):\n",
2021-06-27 11:07:37 +02:00
" cost = cost_function(theta, X, Y)\n",
" logs = [[cost, theta]]\n",
" \n",
" for i in range(max_steps):\n",
2021-06-27 14:09:23 +02:00
" theta = theta - learning_rate * gradient(theta, X, Y)\n",
2021-06-27 11:07:37 +02:00
" next_cost = cost_function(theta, X, Y)\n",
" logs.append([next_cost, theta])\n",
" if abs(cost - next_cost) <= eps:\n",
" break\n",
" cost = next_cost\n",
" return theta, logs\n",
"\n",
2021-06-27 14:09:23 +02:00
"# BGD z momentum\n",
2021-06-27 20:23:26 +02:00
"def momentum(X, Y, theta, cost_function = mean_squared_error, \n",
2021-06-28 10:01:22 +02:00
" learning_rate=0.1, momentum = 0.3, epochs=30):\n",
2021-06-25 01:36:15 +02:00
" cost = cost_function(theta, X, Y)\n",
2021-06-24 13:03:10 +02:00
" logs = [[cost, theta]]\n",
2021-06-27 14:09:23 +02:00
" delta_history = [np.matrix([0]*(degree+1)).T]\n",
" for i in range(epochs):\n",
2021-06-27 20:23:26 +02:00
" delta_history.append(momentum * delta_history[-1] + \n",
" learning_rate * gradient(theta, X, Y))\n",
2021-06-27 14:09:23 +02:00
" theta = theta - delta_history[-1]\n",
2021-06-25 01:36:15 +02:00
" next_cost = cost_function(theta, X, Y)\n",
2021-06-24 13:03:10 +02:00
" logs.append([next_cost, theta])\n",
2021-06-25 01:36:15 +02:00
" cost = next_cost\n",
2021-06-24 13:03:10 +02:00
" return theta, logs\n",
2021-06-25 09:42:03 +02:00
"\n",
2021-06-27 20:23:26 +02:00
"# Mini-batch gradient descend (MBGD)\n",
"def MBGD(X, Y, theta, cost_function = mean_squared_error, \n",
" learning_rate=0.1, epochs=5, batch_size=16):\n",
2021-06-25 09:42:03 +02:00
" cost = cost_function(theta, X, Y)\n",
" logs = [[cost, theta]]\n",
" start, end = 0, batch_size\n",
2021-06-24 13:03:10 +02:00
" \n",
2021-06-25 09:42:03 +02:00
" steps = m / batch_size\n",
" for i in range(epochs):\n",
" zipped_XY = list(zip(X, Y))\n",
" random.shuffle(zipped_XY)\n",
" X_shuffled, Y_shuffled = zip(*zipped_XY)\n",
" X_shuffled = np.concatenate(X_shuffled, axis=0) \n",
" Y_shuffled = np.concatenate(Y_shuffled, axis=0) \n",
" for j in range(int(steps)):\n",
" batch = X_shuffled[start:end,:], Y_shuffled[start:end,:]\n",
2021-06-27 14:09:23 +02:00
" theta = theta - learning_rate * gradient(theta, batch[0], batch[1])\n",
2021-06-25 09:42:03 +02:00
" cost = cost_function(theta, X, Y)\n",
" logs.append([cost, theta])\n",
"\n",
" if start + batch_size < batch_size:\n",
" start += batch_size\n",
" else:\n",
" start = 0\n",
" end = min(start + batch_size, m)\n",
" return theta, logs\n",
"\n",
2021-06-27 20:23:26 +02:00
"# Stochastic gradient descend (SGD)\n",
"def SGD(X, Y, theta, cost_function = mean_squared_error, learning_rate=0.1,\n",
" epochs=5):\n",
" return MBGD(X, Y, theta, cost_function, learning_rate, epochs, 1)"
2021-06-27 17:34:20 +02:00
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## 4. Zestawienie wyników"
]
},
{
"cell_type": "code",
2021-06-28 10:01:22 +02:00
"execution_count": 54,
"metadata": {
"scrolled": true
},
2021-06-27 17:34:20 +02:00
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
2021-06-27 20:23:26 +02:00
"SGD:\n",
2021-06-28 10:01:22 +02:00
"f(x) = 16.32876722693443x^0 + 34.45307103302153x^1 + 45.94300992401012x^2 + 52.620176422095824x^3 + 55.84736037491846^4\n",
2021-06-27 20:23:26 +02:00
"600001 updates\n",
"\n",
2021-06-27 17:34:20 +02:00
"BGD:\n",
2021-06-28 10:01:22 +02:00
"f(x) = -65.64506277392503x^0 + 68.77648241255996x^1 + 114.75349304333463x^2 + 102.20675310280693x^3 + 52.80014652046723^4\n",
2021-06-27 17:34:20 +02:00
"82705 updates\n",
"\n",
"MBGD:\n",
2021-06-28 10:01:22 +02:00
"f(x) = -43.73372717859179x^0 + 44.52908852914625x^1 + 88.75625965596099x^2 + 102.68565372634689x^3 + 96.42646314762048^4\n",
2021-06-27 17:34:20 +02:00
"37501 updates\n",
"\n",
"momentum:\n",
2021-06-28 10:01:22 +02:00
"f(x) = 46.73334506478952x^0 + 46.04136957395274x^1 + 44.48349874963421x^2 + 42.3749456478724x^3 + 39.943624482130815^4\n",
2021-06-27 18:45:49 +02:00
"61 updates\n",
"\n",
"steepest descent:\n",
2021-06-28 10:01:22 +02:00
"f(x) = 13.638648831593645x^0 + 39.092493619820296x^1 + 55.99541645170152x^2 + 66.58927282539155x^3 + 72.5680972400324^4\n",
2021-06-27 17:34:20 +02:00
"61 updates\n",
"\n"
]
}
],
"source": [
2021-06-27 20:23:26 +02:00
"final_theta_SGD, logs_3 = SGD(X, Y, initial_theta, epochs = 60)\n",
"print(\"SGD:\")\n",
"print_theta(final_theta_SGD)\n",
"print(f\"{len(logs_3)} updates\\n\")\n",
"\n",
2021-06-27 14:09:23 +02:00
"final_theta_BGD, logs_1 = BGD(X, Y, initial_theta)\n",
"print(\"BGD:\")\n",
"print_theta(final_theta_BGD)\n",
"print(f\"{len(logs_1)} updates\\n\")\n",
"\n",
"final_theta_MBGD, logs_2 = MBGD(X, Y, initial_theta, epochs = 60)\n",
"print(\"MBGD:\")\n",
"print_theta(final_theta_MBGD)\n",
"print(f\"{len(logs_2)} updates\\n\")\n",
"\n",
2021-06-25 09:42:03 +02:00
"\n",
2021-06-27 14:09:23 +02:00
"final_theta_momentum, logs_4 = momentum(X, Y, initial_theta, epochs = 60)\n",
"print(\"momentum:\")\n",
"print_theta(final_theta_momentum)\n",
2021-06-27 18:45:49 +02:00
"print(f\"{len(logs_4)} updates\\n\")\n",
"\n",
"final_steepest_descent, logs_5 = steepest_descent(X, Y, initial_theta, epochs = 60)\n",
"print(\"steepest descent:\")\n",
"print_theta(final_steepest_descent)\n",
"print(f\"{len(logs_5)} updates\\n\")"
2021-06-25 01:36:15 +02:00
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
2021-06-27 20:23:26 +02:00
"Wniosek:\n",
"Metoda najszybszego spadku oraz momentum cechują się najszybszą zbieżnością do minimum lokalnego danej funkcji. Przeciętnie, dla zadanych parametrów, minimum jest osiągane w ciągu 61 aktualizacji. Podczas gdy SGD wymaga ich ponad 600 tysięcy. "
2021-06-24 13:03:10 +02:00
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
2021-06-28 10:18:23 +02:00
"## 5. Reprezentacja graficzna regresji wielomianowej i funkcji kosztu"
2021-06-24 13:03:10 +02:00
]
},
{
"cell_type": "code",
2021-06-28 10:01:22 +02:00
"execution_count": 52,
2021-06-25 01:36:15 +02:00
"metadata": {
2021-06-27 17:34:20 +02:00
"scrolled": false
2021-06-25 01:36:15 +02:00
},
2021-06-24 13:03:10 +02:00
"outputs": [
2021-06-27 14:09:23 +02:00
{
"name": "stdout",
"output_type": "stream",
"text": [
"BGD:\n"
]
},
{
"data": {
2021-06-27 17:34:20 +02:00
"image/png": "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
2021-06-27 14:09:23 +02:00
"text/plain": [
"<Figure size 720x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
},
2021-06-27 18:45:49 +02:00
{
"data": {
"image/png": "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
"text/plain": [
"<Figure size 720x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
},
2021-06-27 14:09:23 +02:00
{
"name": "stdout",
"output_type": "stream",
"text": [
"MBGD:\n"
]
},
{
"data": {
2021-06-28 10:01:22 +02:00
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAlYAAAEzCAYAAADzdE1rAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAAsTAAALEwEAmpwYAAAwcElEQVR4nO3deXxdZb3v8c+TNB3SQIAwD02qFAWtgkRGQbQoUKwMiqJBUPCGmdajcsHcc5DjiXJFsV6lQHAquEU5ylC0DFJUBBRoAQkIHAokZYa2GoaWNk2e+8dKatruNGmyVvZO8nm/Xn1lZe29n/XwWjT5dq3f+j0hxogkSZIGr6TQE5AkSRopDFaSJEkpMVhJkiSlxGAlSZKUEoOVJElSSgxWkiRJKekzWIUQdgkh/CGE8FgI4dEQwsyu/V8PITwfQnio68/0Hp85P4SwOITwRAjhsCz/AyRJkopF6KuPVQhhB2CHGOMDIYTNgEXA0cCngDdijN9Z7/17ANcA+wA7ArcDu8UYO9KfviRJUvHo84pVjPHFGOMDXduvA48BO23kI0cBv4wxrooxPgMsJglZkiRJI9om1ViFEGqAvYB7u3adFUJ4OITwkxDCll37dgKe7fGx59h4EJMkSRoRxvT3jSGECuA3wKwY42shhMuAbwCx6+t3gZOBkOfjG9xvDCHUA/UAEydO3Pud73znps9ekjTqLXphUa+vbfsm7PIavDYWnqzqx2ABJm8xma0mbJXeBDXiLFq0aGmMcZt8r/UrWIUQykhCVS7GeB1AjPHlHq9fCfy269vngF16fHxn4IX1x4wxNgFNALW1tXHhwoX9mYokSeuomV1Da1vrBvtDJ9z9A9gVOPoYeHL3vscKBJ6+4On0J6kRJYSw4f9wXfrzVGAAfgw8FmO8pMf+HXq87Rjgka7tecDxIYRxIYTJwBTgvoFMXJKkvjROa6S8rHyD/Ucshl3/AS2VcNM7+jfWpMpJa7dzzTlqZtdQcmEJNbNryDXn0pqyRrD+XLE6EPgc0BxCeKhr39eAz4QQ9iS5zdcCnAoQY3w0hHAt8HdgDXCmTwRKkrJSN7UOgJk3z2TZymVr95/dVQ38o/3K6Cxp73Oc8rJyGqc1Akmoqr+pnhXtKwBobWul/qb6dY4n5dNnu4Wh4K1ASVIacs05GhY0MP6pVh7/IawZV8aY51+k5uq9894u7FZdWU3jtMa1oam324vVldW0zGrJavoaJkIIi2KMtfles/O6JGnEqJtaR8usFh4PZwMw5oQToapq7ZWofKomVNEyq2WdK1FL2pbkfW9v+6VuBitJ0rCz0fqn11+Hn/0s2T47CVh1U+uoGFuRd6zXV72+wVg9a6166m2/1M1bgZKkYWX9+qeeqiZUcfM/j+T937qKeyaP4cCT1gAwsWwib7a/2a/xy8vKOem9JzH3b3PXOUZ5WTlNM5qssZK3AiVJI0fDgoa8oQpg+ZvL2KzpKgAuqV2zdn9/QxXAivYVzH9yPk0zmqiurCYQqK6sNlSpX/rdIFSSpGKwsTqnQ5+Gdy6DZzeHGwbRd3pJ2xLqptYZpLTJvGIlSRpWNlbndE5Xi4XLaqGjNJtjSBtjsJIkDSu7brVr3v1vWw7Tn4S3SuHKvQc+fs9+VtKmMlhJkoaNXHOOO565I+9rZ96X/FL75bth6cSBjV81ocpaKg2KNVaSpGGjYUEDkQ2fZp+4Ck5+MNn+wb6bPm7F2Aou/9jlBioNmlesJEnDRm+F6yc8DFusgnt2hgd23PRxx5WOM1QpFQYrSdKwkbeoPMJZ9yWbA7laBayzxqA0GAYrSdKwkGvO5Q1Ahz4N734VXqiA3+w+8PHzdnGXNpE1VpKkopdrznHyjSezumP1Bq996S/J10v3gfZB/FaLRFrbWqm/qR7AW4MaEK9YSZKKUq45x9bf3ppwYeCE607IG6re+SpMXwwrx8AVg2ix0NOK9hU0LGhIZzCNOl6xkiQVXK45R8OCBpa0LWGrCVvx1pq3+rUMzcy/Jl+vei8sG2CLhXw21t1d2hiDlSSpoM743RlcvvDytW0U+ltIXvUmnPi3ZHv2funOyc7rGihvBUqSCibXnFsnVG2KUxdB+RqYvys8vk16c7LzugbDYCVJKpjeGn72pWxN0mkd4Hv7pzef0lBK04wmAGpm1/ikoDaZtwIlSUOqZz3VQEIVwKcfhR3fgOZt4fa3pTe3ztgJQP1N9axoXwHgk4LaJF6xkiQNie6n/E647gRa21oHHKqI/2qxMHs/IKQ2RSZVTqJhQcPaUNXNJwXVXwYrSVLmcs056m+qT6XD+cGt8L6X4JVyyE1NYXJdumurensi0CcF1R8GK0lS5vJdBRqo7qtVl70fVpWlMuTa2qq6qXW9PhHok4LqD4OVJClzaV3tefsy+PgTsKoU5rw/lSEZWzqWucfMXVs/1TitkfKy8nXe45OC6i+DlSQpc1tN2CqVcc65N/nF9Yup8ErF4MermlDFT476yTpF6XVT62ia0UR1ZTWBQHVl9dqrWVJffCpQkpSpXHOO11a9NuhxKlfCyQ8m24NtCFoaSlnzH2t6fb1uap1BSgNisJIkZSbXnOOk60+iI3YMeqwvPgAV7bBgMjy8/eDGqt+7ftDzkfIxWEmSMnHoVYey4JkFqYxVtgZm3ptsXzLIhqDTJk9jzpFzBj8pKQ9rrCRJqTvjd2ekFqoAjn8EdnkNHt0Gbt514ONMmzyNxcsX21FdmfGKlSQpdU2LmtIbLMJX70k2v3MAxEFcEugZ9uyorix4xUqSlLo0aqq6Hb4Ypr4Cz2+WPA2YJjuqK20GK0lS6kpDaWpjffXu5Ov394XVGdxnsaO60mSwkiSlLq2n7vZ+Hj7cAq+NhStqUxlyA3ZUV5oMVpKkVOWac8x/cn4qY3XXVl1RC6+NT2XIDUyfMj2bgTUqWbwuSUpN92LLaawLOHk5fPLv0F6S3AbMSlohUAKvWEmSUpTqYst/hdKYFKw/X5nKkHm1trVmN7hGHYOVJCk1aRWCV70JpzyQbH/ngFSG7FUg2M9KqTFYSZJSk9Ziy2fcD+VrYP6u8Mh2qQzZq0i05YJSY7CSJPUp15yjZnZNnx3L31rz1qCPNb4dzr4v2b74wEEP1y+2XFBaLF6XJG3U+gXpPTuWQ1JXlWad0ucfgm1WwP07wh9rUhuW8rJyJoyZwLKVyzZ4zZYLSovBSpK0UfkK0le0r2DmzTNZuWZlasXqACWd8OWuFgsXHwCEdMatmlDF94/4PsAGTy2Wl5XTOK0xnQNp1PNWoCRpo3q7TbZs5bJUQxXAsY/Brv+Ap7eA63ZPb9yVa1YCyZqATTOaqK6sJhCorqymaUaTawUqNSHGWOg5UFtbGxcuXFjoaUiS8qiZXTM0LQkiPHAF7PUSnH4kXP7+dIevrqymZVZLuoNqVAohLIox5l0LwCtWkqSNapzWSHlZ+Tr7ysvKqZpQlepxDl+chKqXJsJP90x1aMACdQ0Ng5UkaaN63j6DpO/TivYVeYvAB+P8Pydfv3sArCpLdWjAAnUNDYOVJCmvni0WGhY0MH3KdMaUjCGSfgnJB1rh4CXwj/FweQaLLVugrqFisJIkbaC7xUJrWyuRSGtbK5cvvJw1nWsyOV731aof7ANvjEtnzKoJVRaoa8j12W4hhLALcBWwPdAJNMUYvx9C2Ar4FVADtACfijH+o+sz5wOnAB3AOTHGWzOZvSQpE/laLGRxpQrgvS/C9MXwZhn8vxQXW64YW8HSc5emN6DUD/25YrUG+HKMcXdgP+DMEMIewHnAghjjFGBB1/d0vXY88C7gcGBOCKE0i8lLkrIxlIXe59+VfL1ib1g2Mb1xLVZXIfQZrGKML8YYH+jafh14DNgJOAqY2/W2ucDRXdtHAb+MMa6KMT4DLAb2SXnekqQMDVWh95SlcNyjsLoELtk/3bEtVlchbFKNVQihBtgLuBfYLsb4IiThC9i26207Ac/2+NhzXfvWH6s+hLAwhLDw1VdfHcDUJUlZyddiIQvn3p38Ipq7Jzxfmd64FqurUPodrEIIFcBvgFkxxtc29tY8+za4MR9jbIox1sY
2021-06-27 18:45:49 +02:00
"text/plain": [
"<Figure size 720x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
},
{
"data": {
2021-06-28 10:01:22 +02:00
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAnEAAAE9CAYAAABk/zSyAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAAsTAAALEwEAmpwYAAAlTUlEQVR4nO3de5RlZXnn8e9Tl75A01wLxG6wMbYaIEalZfASh4REOo4jJAtiu4x0EmY6EhIvGcfAuNZoMouJxhiVOOJixNA4DkiILpgsUQnekhUEGwW5ibRya+nQLSDdXPpSdZ75Y79Vvav6VPWFc+rUrvp+1jrr7PPuy3n3W/vs+p13n713ZCaSJElqlr5eV0CSJEn7zhAnSZLUQIY4SZKkBjLESZIkNZAhTpIkqYEMcZIkSQ000K0FR8RngTcBmzLzxAnj3gt8BBjKzJ+VsguBc4ER4J2Z+dVSfhJwObAQ+DLwrszMiJgPXAGcBDwGvCUzH9hTvY444ohctmxZJ1ZRkiSpq2699dafZeZQu3FdC3FUweuTVEFrTEQcA/wG8FCt7HhgFXAC8HzgnyLixZk5AlwCrAG+QxXiVgLXUwW+JzLzRRGxCvgw8JY9VWrZsmWsW7fuOa+cJElSt0XEg5ON69rh1Mz8NvB4m1EfA94H1K8yfAZwVWZuz8z7gfXAyRFxNLA4M2/K6qrEVwBn1uZZW4avAU6LiOj8mkiSJM080/qbuIh4M/DTzLx9wqglwMO11xtK2ZIyPLF83DyZOQw8CRzehWpLkiTNON08nDpORBwAvB94Q7vRbcpyivKp5mn33muoDsly7LHH7rGukiRJM9109sT9AnAccHtEPAAsBb4XEc+j6mE7pjbtUuCRUr60TTn1eSJiADiY9odvycxLM3NFZq4YGmr720BJkqRGmbYQl5l3ZOaRmbksM5dRhbBXZua/AdcBqyJifkQcBywHbsnMjcDWiDil/N7tHODassjrgNVl+Czg6+V3c5IkSbNe10JcRFwJ3AS8JCI2RMS5k02bmXcBVwN3A18Bzi9npgKcB3yG6mSHH1OdmQpwGXB4RKwH/hS4oCsrIkmSNAPFXOu8WrFiRXqJEUmS1AQRcWtmrmg3zjs2SJIkNZAhTpIkqYEMcR328OPPcOUtD/HE0zt6XRVJkjSLGeI67O6NW7jwi3fwyJPP9roqkiRpFjPESZIkNZAhTpIkqYEMcZIkSQ1kiJMkSWogQ5wkSVIDGeK6ZI7dCEOSJE0zQ1yHRa8rIEmS5gRDnCRJUgMZ4iRJkhrIECdJktRAhjhJkqQGMsRJkiQ1kCFOkiSpgQxxHRbhRUYkSVL3GeIkSZIayBAnSZLUQIY4SZKkBjLESZIkNZAhTpIkqYEMcV2S2esaSJKk2cwQ12FeYESSJE0HQ5wkSVIDGeIkSZIayBAnSZLUQF0LcRHx2YjYFBF31so+EhE/jIgfRMSXIuKQ2rgLI2J9RNwbEafXyk+KiDvKuIuj3NcqIuZHxBdK+c0Rsaxb6yJJkjTTdLMn7nJg5YSyG4ATM/NlwI+ACwEi4nhgFXBCmedTEdFf5rkEWAMsL4/RZZ4LPJGZLwI+Bny4a2siSZI0w3QtxGXmt4HHJ5R9LTOHy8vvAEvL8BnAVZm5PTPvB9YDJ0fE0cDizLwpMxO4AjizNs/aMnwNcFrMoLvPJ15jRJIkdU8vfxP3B8D1ZXgJ8HBt3IZStqQMTywfN08Jhk8Ch3exvntl5sRISZI0m/UkxEXE+4Fh4POjRW0myynKp5qn3futiYh1EbFu8+bN+1pdSZKkGWfaQ1xErAbeBLytHCKFqoftmNpkS4FHSvnSNuXj5omIAeBgJhy+HZWZl2bmisxcMTQ01KlVkSRJ6plpDXERsRL4M+DNmflMbdR1wKpyxulxVCcw3JKZG4GtEXFK+b3bOcC1tXlWl+GzgK/XQqEkSdKsNtCtBUfElcCpwBERsQH4ANXZqPOBG8o5CN/JzHdk5l0RcTVwN9Vh1vMzc6Qs6jyqM10XUv2GbvR3dJcBn4uI9VQ9cKu6tS6SJEkzTddCXGa+tU3xZVNMfxFwUZvydcCJbcq3AWc/lzpKkiQ1lXds6BIP7EqSpG4yxHWYlxiRJEnTwRAnSZLUQIY4SZKkBjLESZIkNZAhTpIkqYEMcZIkSQ1kiOsSrzAiSZK6yRDXYYHXGJEkSd1niJMkSWogQ5wkSVIDGeIkSZIayBAnSZLUQIY4SZKkBjLEdUmmFxmRJEndY4jrNK8wIkmSpoEhTpIkqYEMcZIkSQ1kiJMkSWogQ5wkSVIDGeIkSZIayBDXJV5gRJIkdZMhrsO8wogkSZoOhjhJkqQGMsRJkiQ1kCFOkiSpgQxxkiRJDWSIkyRJaqCuhbiI+GxEbIqIO2tlh0XEDRFxX3k+tDbuwohYHxH3RsTptfKTIuKOMu7iiIhSPj8ivlDKb46IZd1al/2RXmNEkiR1UTd74i4HVk4ouwC4MTOXAzeW10TE8cAq4IQyz6cior/McwmwBlheHqPLPBd4IjNfBHwM+HDX1mQflIwpSZLUVV0LcZn5beDxCcVnAGvL8FrgzFr5VZm5PTPvB9YDJ0fE0cDizLwpMxO4YsI8o8u6BjgtTFCSJGmOmO7fxB2VmRsByvORpXwJ8HBtug2lbEkZnlg+bp7MHAaeBA5v96YRsSYi1kXEus2bN3doVSRJknpnppzY0K4HLacon2qe3QszL83MFZm5YmhoaD+rKEmSNHNMd4h7tBwipTxvKuUbgGNq0y0FHinlS9uUj5snIgaAg9n98K0kSdKsNN0h7jpgdRleDVxbK19Vzjg9juoEhlvKIdetEXFK+b3bORPmGV3WWcDXy+/mJEmSZr2Bbi04Iq4ETgWOiIgNwAeADwFXR8S5wEPA2QCZeVdEXA3cDQwD52fmSFnUeVRnui4Eri8PgMuAz0XEeqoeuFXdWpf9Y56UJEnd07UQl5lvnWTUaZNMfxFwUZvydcCJbcq3UULgTOLpsZIkaTrMlBMbJEmStA8McZIkSQ1kiJMkSWogQ5wkSVIDGeIkSZIayBDXJV6xTpIkdZMhrsPCa4xIkqRpYIiTJElqIEOcJElSAxniJEmSGsgQJ0mS1ECGOEmSpAYyxHWJVxiRJEndZIjrsMBrjEiSpO4zxEmSJDWQIU6SJKmBDHGSJEkNZIiTJElqIEOcJElSAxniuiS9xogkSeoiQ1yHhVcYkSRJ08AQJ0mS1ECGOEmSpAYyxEmSJDWQIU6SJKmBDHGSJEkNZIjrkvQaI5IkqYt6EuIi4j0RcVdE3BkRV0bEgog4LCJuiIj7yvOhtekvjIj1EXFvRJxeKz8pIu4o4y6O6P0FPnpeAUmSNCdMe4iLiCXAO4EVmXki0A+sAi4AbszM5cCN5TURcXwZfwKwEvhURPSXxV0CrAGWl8fKaVwVSZKknunV4dQBYGFEDAAHAI8AZwBry/i1wJll+Azgqszcnpn3A+uBkyPiaGBxZt6U1bHLK2rzSJIkzWrTHuIy86fAXwMPARuBJzPza8BRmbmxTLMROLLMsgR4uLaIDaVsSRmeWC5JkjTr9eJw6qFUvWvHAc8HDoyI351qljZlOUV5u/dcExHrImLd5s2b97XKkiRJM04vDqf+OnB/Zm7OzJ3AF4HXAI+WQ6SU501l+g3AMbX5l1Idft1QhieW7yYzL83MFZm5YmhoqKMrI0mS1Au9CHEPAadExAHlbNLTgHuA64DVZZrVwLVl+DpgVUTMj4jjqE5guKUcct0aEaeU5ZxTm6fnvMCIJEnqpoHpfsPMvDkirgG+BwwD3wcuBRYBV0fEuVRB7+wy/V0RcTVwd5n+/MwcKYs7D7gcWAhcXx695TVGJEnSNJj2EAeQmR8APjCheDtVr1y76S8CLmpTvg44seMVlCRJmuG8Y4MkSVIDGeIkSZIayBAnSZLUQIY4SZKkBjLEdUl6jRFJktRFhrgOC68xIkmSpoEhTpIkqYEMcZIkSQ20VyEuIj63N2WSJEmaHnvbE3dC/UVE9AMndb46kiRJ2htThriIuDAitgIvi4gt5bEV2MQMutm8JEnSXDNliMv
2021-06-27 14:09:23 +02:00
"text/plain": [
"<Figure size 720x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"SGD:\n"
]
},
2021-06-24 13:03:10 +02:00
{
"data": {
2021-06-28 10:01:22 +02:00
"image/png": "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
2021-06-27 18:45:49 +02:00
"text/plain": [
"<Figure size 720x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
},
{
"data": {
2021-06-28 10:01:22 +02:00
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAnEAAAE9CAYAAABk/zSyAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAAsTAAALEwEAmpwYAAArNklEQVR4nO3de5xcZZ3n8c+vu3MDEkigEyAJJGAAuSiYDHLTQQHBGRWc9RJXh6jMoC6z3tZVorvqzA4zOjqO4gWH9QI4CjIMLNFREcG7CDYQhAAh4ZY0CaQJBEKAJN312z/qdChCVXWTdHf1CZ/361WvOvWc23Oe7q7+1vOcUycyE0mSJJVLW6srIEmSpOfPECdJklRChjhJkqQSMsRJkiSVkCFOkiSphAxxkiRJJdQxXBuOiG8BrwPWZOahW837CPA5oDMzHy7KFgJnAH3A+zPzqqJ8LnABMAH4EfCBzMyIGAdcBMwF1gJvzcz7BqrXHnvskbNmzRqKQ5QkSRpWN95448OZ2Vlv3rCFOKrB6ytUg9YWETETOAlYUVN2MDAfOATYG/hZRByQmX3AecCZwO+phrhTgB9TDXyPZuaLImI+8FngrQNVatasWXR1dW33wUmSJA23iLi/0bxhG07NzF8Bj9SZ9S/AR4Habxk+FbgkMzdm5r3AcuDIiNgLmJSZ12X1W4kvAk6rWefCYvoy4ISIiKE/EkmSpNFnRM+Ji4g3AA9k5i1bzZoOrKx53V2UTS+mty5/1jqZ2Qs8Buw+DNWWJEkadYZzOPVZImIn4BPAa+rNrlOWTcqbrVNv32dSHZJln332GbCukiRJo91I9sTtD8wGbomI+4AZwE0RsSfVHraZNcvOAFYV5TPqlFO7TkR0ALtSf/iWzDw/M+dl5rzOzrrnBkqSJJXKiIW4zLw1M6dm5qzMnEU1hL0sMx8EFgHzI2JcRMwG5gA3ZOZqYH1EHFWc73Y6cGWxyUXAgmL6TcC1xXlzkiRJO7xhC3ERcTFwHXBgRHRHxBmNls3MJcClwO3AT4CziitTAd4HfIPqxQ53U70yFeCbwO4RsRz4MHD2sByIJEnSKBQvtM6refPmpV8xIkmSyiAibszMefXmeccGSZKkEjLESZIklZAhboitfORJLr5hBY9u2NTqqkiSpB2YIW6I3b76cRZefiurHnuq1VWRJEk7MEOcJElSCRniJEmSSsgQJ0mSVEKGOEmSpBIyxEmSJJWQIU6SJKmEDHGSJEklZIiTJEkqIUOcJElSCRniJEmSSsgQJ0mSVEKGOEmSpBIyxEmSJJWQIU6SJKmEDHGSJEklZIiTJEkqIUOcJElSCRniJEmSSsgQJ0mSVEKGOEmSpBIyxEmSJJWQIU6SJKmEDHGSJEklZIiTJEkqoWELcRHxrYhYExG31ZR9LiLujIg/RsQVEbFbzbyFEbE8IpZGxMk15XMj4tZi3rkREUX5uIj4flF+fUTMGq5jkSRJGm2GsyfuAuCUrcquBg7NzJcAdwELASLiYGA+cEixztcior1Y5zzgTGBO8ejf5hnAo5n5IuBfgM8O25FIkiSNMsMW4jLzV8AjW5X9NDN7i5e/B2YU06cCl2Tmxsy8F1gOHBkRewGTMvO6zEzgIuC0mnUuLKYvA07o76WTJEna0bXynLh3Az8upqcDK2vmdRdl04vprcuftU4RDB8Ddh/G+kqSJI0aLQlxEfEJoBf4bn9RncWySXmzdert78yI6IqIrp6enudbXUmSpFFnxENcRCwAXge8vRgihWoP28yaxWYAq4ryGXXKn7VORHQAu7LV8G2/zDw/M+dl5rzOzs6hOhRJkqSWGdEQFxGnAB8D3pCZT9bMWgTML644nU31AoYbMnM1sD4ijirOdzsduLJmnQXF9JuAa2tCoSRJ0g6tY7g2HBEXA8cDe0REN/ApqlejjgOuLq5B+H1mvjczl0TEpcDtVIdZz8rMvmJT76N6pesEqufQ9Z9H903gOxGxnGoP3PzhOhZJkqTRZthCXGa+rU7xN5ssfw5wTp3yLuDQOuVPA2/enjpKkiSVlXdskCRJKiFDnCRJUgkZ4iRJkkrIECdJklRChjhJkqQSMsRJkiSVkCFOkiSphAxxkiRJJWSIkyRJKiFDnCRJUgkZ4iRJkkrIECdJklRChjhJkqQSMsRJkiSVkCFOkiSphAxxkiRJJWSIkyRJKiFDnCRJUgkZ4iRJkkrIECdJklRChjhJkqQSMsRJkiSVkCFOkiSphAxxkiRJJWSIkyRJKiFDnCRJUgkZ4iRJkkrIECdJklRCwxbiIuJbEbEmIm6rKZsSEVdHxLLieXLNvIURsTwilkbEyTXlcyPi1mLeuRERRfm4iPh+UX59RMwarmORJEkabYazJ+4C4JStys4GrsnMOcA1xWsi4mBgPnBIsc7XIqK9WOc84ExgTvHo3+YZwKOZ+SLgX4DPDtuRSJIkjTLDFuIy81fAI1sVnwpcWExfCJxWU35JZm7MzHuB5cCREbEXMCkzr8vMBC7aap3+bV0GnNDfSydJkrSjG+lz4qZl5mqA4nlqUT4dWFmzXHdRNr2Y3rr8WetkZi/wGLD7sNVckiRpFBktFzbU60HLJuXN1nnuxiPOjIiuiOjq6enZxipKkiSNHiMd4h4qhkgpntcU5d3AzJrlZgCrivIZdcqftU5EdAC78tzhWwAy8/zMnJeZ8zo7O4foUCRJklpnpEPcImBBMb0AuLKmfH5xxelsqhcw3FAMua6PiKOK891O32qd/m29Cbi2OG9OkiRph9cxXBuOiIuB44E9IqIb+BTwGeDSiDgDWAG8GSAzl0TEpcDtQC9wVmb2FZt6H9UrXScAPy4eAN8EvhMRy6n2wM0frmORJEkabYYtxGXm2xrMOqHB8ucA59Qp7wIOrVP+NEUIlCRJeqEZLRc2SJIk6XkwxEmSJJWQIU6SJKmEDHGSJEklZIiTJEkqIUOcJElSCRniJEmSSsgQJ0mSVEKGOEmSpBIyxEmSJJWQIU6SJKmEDHGSJEklZIiTJEkqIUOcJElSCRniJEmSSsgQJ0mSVEKGOEmSpBIyxEmSJJWQIU6SJKmEDHGSJEklZIiTJEkqIUOcJElSCRniJEmSSsgQJ0mSVEKGOEmSpBIyxEmSJJWQIU6SJKmEDHGSJEkl1JIQFxEfioglEXFbRFwcEeMjYkpEXB0Ry4rnyTXLL4yI5RGxNCJOrimfGxG3FvPOjYhoxfFIkiSNtBEPcRExHXg/MC8zDwXagfnA2cA1mTkHuKZ4TUQcXMw/BDgF+FpEtBebOw84E5hTPE4ZwUORJElqmVYNp3YAEyKiA9gJWAWcClxYzL8QOK2YPhW4JDM3Zua9wHLgyIjYC5iUmddlZgIX1awjSZK0QxvxEJeZDwCfB1YAq4HHMvOnwLTMXF0ssxqYWqwyHVhZs4nuomx6Mb11uSRJ0g6vFcOpk6n2rs0G9gZ2joh3NFulTlk2Ka+3zzMjoisiunp6ep5vlSVJkkadVgynngjcm5k9mbkZuBw4BnioGCKleF5TLN8NzKxZfwbV4dfuYnrr8ufIzPMzc15mzuvs7BzSg5EkSWqFVoS4FcBREbFTcTXpCcAdwCJgQbHMAuDKYnoRMD8ixkXEbKoXMNxQDLmuj4ijiu2cXrOOJEnSDq1jpHeYmddHxGXATUAvcDNwPrALcGlEnEE16L25WH5JRFwK3F4sf1Zm9hWbex9wATAB+HHxkCRJ2uGNeIgDyMxPAZ/aqngj1V65esufA5xTp7wLOHTIKyhJkjTKeccGSZKkEjLESZIklZAhTpIkqYQMcZIkSSVkiJMkSSohQ5wkSVIJGeIkSZJKaFAhLiK+M5gySZIkjYzB9sQdUvsiItqBuUNfHUmSJA1G0xAXEQsjYj3wkoh4vHisp3pzeu9TKkmS1CJNQ1xm/mNmTgQ+l5mTisfEzNw9MxeOUB0lSZK0lcEOp/4wInYGiIh3RMQXImLfYayXJEmSmhhsiDsPeDIiXgp8FLgfuGjYaiVJkqSmBhviejMzgVOBL2Xml4CJw1ctSZIkNdMxyOXWR8RC4C+BVxRXp44ZvmqVX2arayBJknZ
2021-06-27 14:09:23 +02:00
"text/plain": [
"<Figure size 720x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Momentum:\n"
]
},
{
"data": {
2021-06-28 10:01:22 +02:00
"image/png": "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
2021-06-24 13:03:10 +02:00
"text/plain": [
"<Figure size 720x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
2021-06-27 18:45:49 +02:00
},
{
"data": {
2021-06-28 10:01:22 +02:00
"image/png": "iVBORw0KGgoAAAANSUhEUgAAAnEAAAE9CAYAAABk/zSyAAAAOXRFWHRTb2Z0d2FyZQBNYXRwbG90bGliIHZlcnNpb24zLjMuNCwgaHR0cHM6Ly9tYXRwbG90bGliLm9yZy8QVMy6AAAACXBIWXMAAAsTAAALEwEAmpwYAAAnWUlEQVR4nO3de5RedX3v8fc3cx/IzIRkMoEEDWDUAipKpAR7saKVth6hHtS4VNCyTo4eekTbHgv1D3u6Fkc9nlalLbhYogRLRUr1QFtviPXSIwJBUeQmAQQikIRLwkBIyOV7/nj2hIfJk7k++7lM3q+1Zj37+e3bb/8yyfrkt397/yIzkSRJUnuZ1+wKSJIkafoMcZIkSW3IECdJktSGDHGSJEltyBAnSZLUhgxxkiRJbaizrANHxOeBNwGbMvPYcev+DPgkMJyZjxZl5wFnAbuBD2TmN4vy44FLgT7ga8A5mZkR0QNcBhwPPAa8PTN/OVm9Fi1alMuXL6/HJUqSJJXq5ptvfjQzh2utKy3EUQlef0claO0VEYcDbwAeqCo7GlgNHAMcBnw7Il6cmbuBi4A1wI+ohLhTgK9TCXxPZOaLImI18Ang7ZNVavny5axbt27WFydJklS2iLh/f+tKu52amd8HHq+x6lPAh4HqtwyfClyRmTsy8z5gPXBCRBwKDGTm9Vl5K/FlwGlV+6wtlq8CTo6IqP+VSJIktZ6GjomLiDcDv8rMn45btRR4sOr7hqJsabE8vvx5+2TmLmArsLCEakuSJLWcMm+nPk9E9AMfAX631uoaZTlB+UT71Dr3Giq3ZHnBC14waV0lSZJaXSN74o4CjgB+GhG/BJYBP46IJVR62A6v2nYZ8FBRvqxGOdX7REQnMEjt27dk5sWZuTIzVw4P1xwbKEmS1FYaFuIy89bMXJyZyzNzOZUQ9qrMfAS4BlgdET0RcQSwArgxMx8GRiPixGK82xnA1cUhrwHOLJZPB75TjJuTJEma80oLcRHxJeB64CURsSEiztrftpl5G3AlcDvwDeDs4slUgPcDn6PysMM9VJ5MBbgEWBgR64E/Ac4t5UIkSZJaUBxonVcrV65MXzEiSZLaQUTcnJkra61zxgZJkqQ2ZIiTJElqQ4a4Onvw8W1cfsP9bN22s9lVkSRJc5ghrs7ufGSUj3z15zzw+LZmV0WSJM1hhrg6G+rvAmDLM882uSaSJGkuM8TV2VBfEeK8nSpJkkpkiKuzwb09cYY4SZJUHkNcnQ0WPXFbt3k7VZIklccQV2c9nR30d3d4O1WSJJXKEFeCob4ub6dKkqRSGeJKMNjfbU+cJEkqlSGuBEN9XWz1FSOSJKlEhrgSDPV32RMnSZJKZYgrwVC/Y+IkSVK5DHElGOzrZuu2nWRms6siSZLmKENcCYb6u3h29x6e2bm72VWRJElzlCGuBE69JUmSymaIK8FQvyFOkiSVyxBXgsG+bgC2+JoRSZJUEkNcCcZ64rbaEydJkkpiiCvB3tupvmZEkiSVxBBXgqGx26n2xEmSpJIY4krQ2zWP7s55jomTJEmlMcSVICIq86faEydJkkpiiCuJ86dKkqQyGeJKMtTX7e1USZJUGkNcSQbtiZMkSSUyxJVkqK+Lrb5iRJIklaS0EBcRn4+ITRHx86qyT0bEnRHxs4j4akQMVa07LyLWR8RdEfHGqvLjI+LWYt0FERFFeU9EfLkovyEilpd1LTPhmDhJklSmMnviLgVOGVd2LXBsZr4c+AVwHkBEHA2sBo4p9rkwIjqKfS4C1gArip+xY54FPJGZLwI+BXyitCuZgaH+bp7ZuZvtO3c3uyqSJGkOKi3EZeb3gcfHlX0rM3cVX38ELCuWTwWuyMwdmXkfsB44ISIOBQYy8/rMTOAy4LSqfdYWy1cBJ4/10rWCwb7KrA1PektVkiSVoJlj4v4I+HqxvBR4sGrdhqJsabE8vvx5+xTBcCuwsMT6TotTb0mSpDI1JcRFxEeAXcDlY0U1NssJyifap9b51kTEuohYt3nz5ulWd0acekuSJJWp4SEuIs4E3gS8s7hFCpUetsOrNlsGPFSUL6tR/rx9IqITGGTc7dsxmXlxZq7MzJXDw8P1upQJ7e2J2+a74iRJUv01NMRFxCnAnwNvzsxtVauuAVYXT5weQeUBhhsz82FgNCJOLMa7nQFcXbXPmcXy6cB3qkJh042NifN2qiRJKkNnWQeOiC8BrwUWRcQG4KNUnkbtAa4tnkH4UWa+LzNvi4grgdup3GY9OzPHHut8P5UnXfuojKEbG0d3CfDFiFhPpQdudVnXMhNjPXHOnypJkspQWojLzHfUKL5kgu3PB86vUb4OOLZG+XbgrbOpY5kO7umkY1449ZYkSSqFMzaUJCIY6vOFv5IkqRyGuBIN9nc5Jk6SJJXCEFeiob4ux8RJkqRSGOJKNNTf7Zg4SZJUCkNciRwTJ0mSymKIK9Fgv7dTJUlSOQxxJRrq62Z0xy527t7T7KpIkqQ5xhBXorEX/j7pE6qSJKnODHEl2jt/qiFOkiTVmSGuRHvnT3VcnCRJqjNDXImG+rsB2OprRiRJUp0Z4ko0ZE+cJEkqiSGuRHvHxBniJElSnRniSjS/t4sIH2yQJEn1Z4grUce8YKC3i63bHBMnSZLqyxBXsqH+LnviJElS3RniSub8qZIkqQyGuJIN9nfbEydJkurOEFeyoT7HxEmSpPozxJXMMXGSJKkMhriSDfV1sfWZnezZk82uiiRJmkMMcSUb7O8mE0a372p2VSRJ0hxiiCvZ3qm3nD9VkiTVkSGuZE69JUmSymCIK9neEOfDDZIkqY4McSUb7OsGYIuvGZEkSXVkiCvZWE/cVnviJElSHRniSjbY55g4SZJUf6WFuIj4fERsioifV5UdEhHXRsTdxeeCqnXnRcT6iLgrIt5YVX58RNxarLsgIqIo74mILxflN0TE8rKuZTa6OuZxcE+nIU6SJNVVmT1xlwKnjCs7F7guM1cA1xXfiYijgdXAMcU+F0ZER7HPRcAaYEXxM3bMs4AnMvNFwKeAT5R2JbM02NflK0YkSVJdlRbiMvP7wOPjik8F1hbLa4HTqsqvyMwdmXkfsB44ISIOBQYy8/rMTOCycfuMHesq4OSxXrpWM9TfxVZ74iRJUh01ekzcSGY+DFB8Li7KlwIPVm23oShbWiyPL3/ePpm5C9gKLCyt5rPg/KmSJKneWuXBhlo9aDlB+UT77HvwiDURsS4i1m3evHmGVZy5ob5uXzEiSZLqqtEhbmNxi5Tic1NRvgE4vGq7ZcBDRfmyGuXP2yciOoFB9r19C0BmXpyZKzNz5fDwcJ0uZeoG+7t8xYgkSaqrRoe4a4Azi+UzgaurylcXT5weQeUBhhuLW66jEXFiMd7tjHH7jB3rdOA7xbi5ljPU18WWbTtp0epJkqQ21FnWgSPiS8BrgUURsQH4KPBx4MqIOAt4AHgrQGbeFhFXArcDu4CzM3N3caj3U3nStQ/4evEDcAnwxYhYT6UHbnVZ1zJbQ/1d7NqTPP3sbg7uKa3JJUnSAaS0RJGZ79jPqpP3s/35wPk1ytcBx9Yo304RAlvdUNXUW4Y4SZJUD63yYMOcNtjvrA2SJKm+DHENMNTn/KmSJKm+DHENMNRfuZ36hK8ZkSRJdWKIa4Ahb6dKkqQ6M8Q1wKC3UyVJUp0Z4hqgt6uD3q55ztogSZLqxhDXIJWpt+yJkyRJ9WGIa5Ch/i62eDtVkiTViSGuQQb7uthqT5wkSaoTQ1yDVHriHBMnSZLqwxDXII6JkyRJ9WSIa5CxMXGZ2eyqSJKkOcAQ1yCD/V08u2sP23fuaXZVJEnSHGCIa5ChvsrUW46LkyRJ9WCIaxCn3pIkSfVkiGuQoT5DnCRJqh9DXIMM9o/Nn+rtVEmSNHuGuAYZ6i/GxNkTJ0mS6sAQ1yALxsbEOfWWJEmqA0Ncg/R1ddDdMc+eOEmSVBeGuAaJCAb7uxwTJ0mS6sIQ10BDfV32xEmSpLowxDXQUL8hTpIk1YchroEG+7p9sEGSJNWFIa6Bhvq72LrNMXGSJGn2DHENNNTXZU+cJEmqC0NcAw3
2021-06-27 18:45:49 +02:00
"text/plain": [
"<Figure size 720x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
},
{
"name": "stdout",
"output_type": "stream",
"text": [
"Metoda najszybszego spadku:\n"
]
},
{
"data": {
"image/png": "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
"text/plain": [
"<Figure size 720x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
},
{
"data": {
"image/png": "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
"text/plain": [
"<Figure size 720x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
2021-06-24 13:03:10 +02:00
}
],
"source": [
"def plot_polynomial_regression(theta):\n",
" fig = plt.figure(figsize=(10,5))\n",
2021-06-25 01:36:15 +02:00
" Y_plot = polynomial_regression(theta, X_plot).tolist()\n",
2021-06-24 13:03:10 +02:00
" chart = fig.add_subplot()\n",
2021-06-27 11:07:37 +02:00
" chart.plot(data[\"Height\"], Y ,\"go\")\n",
2021-06-24 13:03:10 +02:00
" chart.plot(X_plot, Y_plot, color=\"red\", lw=2, label=f\"degree {len(theta)}\")\n",
2021-06-27 11:07:37 +02:00
" plt.ylim([0,250])\n",
2021-06-24 13:03:10 +02:00
" plt.show()\n",
2021-06-27 18:45:49 +02:00
" \n",
"\n",
"def plot_cost_function(logs):\n",
" no_iter = no_iter = [x for x in range(1, len(logs) + 1)]\n",
" all_cost = [row[0] for row in logs]\n",
" \n",
" fig = plt.figure(figsize=(10,5)) \n",
" plt.plot(no_iter, all_cost)\n",
" plt.xlabel('No. of iterations')\n",
" plt.ylabel('Cost')\n",
" plt.show()\n",
" \n",
2021-06-27 14:09:23 +02:00
"\n",
"print(\"BGD:\")\n",
"plot_polynomial_regression(final_theta_BGD)\n",
2021-06-27 18:45:49 +02:00
"plot_cost_function(logs_1)\n",
2021-06-27 14:09:23 +02:00
"print(\"MBGD:\")\n",
"plot_polynomial_regression(final_theta_MBGD)\n",
2021-06-27 18:45:49 +02:00
"plot_cost_function(logs_2)\n",
2021-06-27 14:09:23 +02:00
"print(\"SGD:\")\n",
"plot_polynomial_regression(final_theta_SGD)\n",
2021-06-27 18:45:49 +02:00
"plot_cost_function(logs_3)\n",
2021-06-27 14:09:23 +02:00
"print(\"Momentum:\")\n",
2021-06-27 18:45:49 +02:00
"plot_polynomial_regression(final_theta_momentum)\n",
"plot_cost_function(logs_4)\n",
"print(\"Metoda najszybszego spadku:\")\n",
"plot_polynomial_regression(final_steepest_descent)\n",
"plot_cost_function(logs_5)"
]
},
2021-06-28 00:02:15 +02:00
{
"cell_type": "markdown",
"metadata": {},
"source": [
"### Wizualizacja metody najszybszego spadku dla wielomianu drugiego stopnia"
]
},
{
"cell_type": "code",
2021-06-28 10:01:22 +02:00
"execution_count": 11,
2021-06-28 00:02:15 +02:00
"metadata": {},
"outputs": [],
"source": [
"degree = 1\n",
"initial_theta = np.matrix([0] * (degree + 1)).reshape(degree + 1, 1)\n",
"m, n_plus_1 = data_matrix.shape\n",
"n = n_plus_1 - 1\n",
"X = (np.ones((m, 1)))\n",
"\n",
"for i in range(1, degree + 1):\n",
" Xn = np.power(data_matrix[:, 0:n], i)\n",
" Xn /= np.amax(Xn, axis=0)\n",
" X = np.concatenate((X, Xn), axis=1)\n",
"\n",
"X = np.matrix(X).reshape(m, degree * n + 1)\n",
"Y = np.matrix(data_matrix[:, -1])\n",
"\n",
"\n",
"steepest_descent_deg_2, logs_deg_2 = steepest_descent(X, Y, initial_theta, epochs = 60)\n",
"\n"
]
},
{
"cell_type": "code",
2021-06-28 10:01:22 +02:00
"execution_count": 12,
2021-06-28 00:02:15 +02:00
"metadata": {},
"outputs": [],
"source": [
"cost_history = [row[0] for row in logs_deg_2]\n",
"all_thetas = [row[1] for row in logs_deg_2]\n",
"theta0_history = [row[0].item() for row in all_thetas]\n",
"theta1_history = [row[1].item() for row in all_thetas]\n",
"\n",
"cost_history = np.array(cost_history)\n",
"theta0_history = np.array(theta0_history)\n",
"theta1_history = np.array(theta1_history)"
]
},
{
"cell_type": "code",
2021-06-28 10:01:22 +02:00
"execution_count": 13,
2021-06-28 00:02:15 +02:00
"metadata": {},
"outputs": [],
"source": [
"theta0_vals = theta0_history\n",
"theta1_vals = theta1_history\n",
"J_vals = np.zeros((len(theta0_vals), len(theta1_vals)))\n",
"\n",
"c1=0\n",
"c2=0\n",
"pom = 0\n",
"for i in theta0_vals:\n",
" for j in theta1_vals:\n",
" t = np.array([i, j])\n",
" J_vals[c1][c2] = cost_history[pom]\n",
" c2=c2+1\n",
" c1=c1+1\n",
" pom += 1\n",
" c2=0"
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"import plotly.graph_objects as go\n",
"\n",
"fig = go.Figure(data=[go.Surface(x=theta0_vals, y=theta1_vals, z=J_vals)])\n",
"fig.update_layout(title='Loss function for different thetas', autosize=True,\n",
" width=600, height=600, xaxis_title='theta0', \n",
" yaxis_title='theta1')\n",
"fig.show()"
]
},
2021-06-24 13:03:10 +02:00
{
"cell_type": "markdown",
"metadata": {},
"source": [
2021-06-28 10:18:23 +02:00
"## 6. Regresja wielomianowa z wykorzystaniem gotowych bibliotek"
2021-06-24 13:03:10 +02:00
]
},
{
"cell_type": "code",
2021-06-27 20:23:26 +02:00
"execution_count": 10,
2021-06-24 13:03:10 +02:00
"metadata": {},
"outputs": [],
"source": [
"from sklearn.preprocessing import PolynomialFeatures, StandardScaler\n",
"from sklearn.pipeline import make_pipeline\n",
"from sklearn.linear_model import Ridge, LinearRegression"
]
},
{
"cell_type": "code",
2021-06-27 20:23:26 +02:00
"execution_count": 11,
2021-06-24 13:03:10 +02:00
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
2021-06-27 11:07:37 +02:00
"Pipeline(steps=[('polynomialfeatures', PolynomialFeatures(degree=4)),\n",
2021-06-24 13:03:10 +02:00
" ('linearregression', LinearRegression())])"
]
},
2021-06-27 20:23:26 +02:00
"execution_count": 11,
2021-06-24 13:03:10 +02:00
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"model = make_pipeline(PolynomialFeatures(degree=degree, include_bias=True), \n",
" LinearRegression())\n",
2021-06-25 01:36:15 +02:00
"model.fit(data[[\"Height\"]],Y)"
2021-06-24 13:03:10 +02:00
]
},
{
"cell_type": "code",
2021-06-27 20:23:26 +02:00
"execution_count": 12,
2021-06-24 13:03:10 +02:00
"metadata": {},
"outputs": [
{
"data": {
"text/plain": [
2021-06-27 11:07:37 +02:00
"4"
2021-06-24 13:03:10 +02:00
]
},
2021-06-27 20:23:26 +02:00
"execution_count": 12,
2021-06-24 13:03:10 +02:00
"metadata": {},
"output_type": "execute_result"
},
{
"data": {
2021-06-27 17:34:20 +02:00
"image/png": "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
2021-06-24 13:03:10 +02:00
"text/plain": [
"<Figure size 720x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
2021-06-27 11:07:37 +02:00
"Y_plot_2 = model.predict([[x] for x in X_plot])\n",
2021-06-24 13:03:10 +02:00
"\n",
"fig = plt.figure(figsize=(10,5))\n",
"chart = fig.add_subplot()\n",
2021-06-25 01:36:15 +02:00
"chart.plot(data[\"Height\"], Y ,\"go\")\n",
2021-06-27 11:07:37 +02:00
"chart.plot(X_plot, Y_plot_2, color=\"red\", lw=2, label=f\"degree {degree}\")\n",
2021-06-26 17:59:05 +02:00
"plt.ylim([0,250])\n",
2021-06-27 11:07:37 +02:00
"plt.xlim([0,100])\n",
2021-06-24 13:03:10 +02:00
"degree"
]
2021-06-27 14:09:23 +02:00
},
{
"cell_type": "code",
2021-06-27 20:23:26 +02:00
"execution_count": 13,
2021-06-27 14:09:23 +02:00
"metadata": {},
"outputs": [
{
"data": {
2021-06-27 17:34:20 +02:00
"image/png": "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
2021-06-27 14:09:23 +02:00
"text/plain": [
"<Figure size 720x360 with 1 Axes>"
]
},
"metadata": {
"needs_background": "light"
},
"output_type": "display_data"
}
],
"source": [
"polyfit_theta = np.polyfit(data[\"Height\"]/data[\"Height\"].max(), Y, degree)\n",
"plot_polynomial_regression(polyfit_theta)"
]
2021-06-24 11:07:41 +02:00
}
],
"metadata": {
"kernelspec": {
"display_name": "Python 3",
"language": "python",
"name": "python3"
},
"language_info": {
"codemirror_mode": {
"name": "ipython",
"version": 3
},
"file_extension": ".py",
"mimetype": "text/x-python",
"name": "python",
"nbconvert_exporter": "python",
"pygments_lexer": "ipython3",
2021-06-27 17:34:20 +02:00
"version": "3.8.8"
2021-06-24 11:07:41 +02:00
}
},
"nbformat": 4,
"nbformat_minor": 4
}
