AB#1008732: Create and understand classification models in machine learning

Author: gitName

learn-pr/azure/understand-classification-machine-learning/notebooks/6-3-exercise-build-regression-model.ipynb

@@ -56,12 +56,31 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "import graphing # custom graphing code. See our GitHub repo for details\n",
+    "import matplotlib.pyplot as plt\n",
+    "\n",
+    "# Create subplots\n",
+    "fig, axs = plt.subplots(2, 2)\n",
+    "\n",
+    "# Create box plots\n",
+    "axs[0, 0].boxplot([dataset[\"avalanche\"], dataset[\"surface_hoar\"]])\n",
+    "axs[0, 0].set_xticklabels(['Avalanche', 'Surface Hoar'])\n",
+    "axs[0, 0].set_ylabel('Values')\n",
+    "\n",
+    "axs[0, 1].boxplot([dataset[\"avalanche\"], dataset[\"fresh_thickness\"]])\n",
+    "axs[0, 1].set_xticklabels(['Avalanche', 'Fresh Thickness'])\n",
+    "axs[0, 1].set_ylabel('Values')\n",
+    "\n",
+    "axs[1, 0].boxplot([dataset[\"avalanche\"], dataset[\"weak_layers\"]])\n",
+    "axs[1, 0].set_xticklabels(['Avalanche', 'Weak Layers'])\n",
+    "axs[1, 0].set_ylabel('Values')\n",
+    "\n",
+    "axs[1, 1].boxplot([dataset[\"avalanche\"], dataset[\"no_visitors\"]])\n",
+    "axs[1, 1].set_xticklabels(['Avalanche', 'Visitors'])\n",
+    "axs[1, 1].set_ylabel('Values')\n",
     "\n",
-    "graphing.box_and_whisker(dataset, label_x=\"avalanche\", label_y=\"surface_hoar\", show=True)\n",
-    "graphing.box_and_whisker(dataset, label_x=\"avalanche\", label_y=\"fresh_thickness\", show=True)\n",
-    "graphing.box_and_whisker(dataset, label_x=\"avalanche\", label_y=\"weak_layers\", show=True)\n",
-    "graphing.box_and_whisker(dataset, label_x=\"avalanche\", label_y=\"no_visitors\")"
+    "plt.tight_layout()\n",
+    "\n",
+    "plt.show()"
    ]
   },
   {
@@ -196,14 +215,24 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "# plot the model\n",
-    "# Show a graph of the result\n",
-    "predict = lambda x: model.predict(pandas.DataFrame({\"weak_layers\": x}))\n",
-    "\n",
-    "graphing.line_2D([(\"Model\", predict)],\n",
-    "                 x_range=[-20,40],\n",
-    "                 label_x=\"weak_layers\", \n",
-    "                 label_y=\"estimated probability of an avalanche\")"
+    "import matplotlib.pyplot as plt\n",
+    "import pandas as pd\n",
+    "\n",
+    "predict = lambda x: model.predict(pd.DataFrame({\"weak_layers\": x}))\n",
+    "\n",
+    "# Generate x values\n",
+    "x_values = range(-20, 41)\n",
+    "y_values = predict(x_values)\n",
+    "\n",
+    "# Create the plot\n",
+    "plt.plot(x_values, y_values, label=\"Model\")\n",
+    "\n",
+    "# Add labels\n",
+    "plt.xlabel(\"weak_layers\")\n",
+    "plt.ylabel(\"estimated probability of an avalanche\")\n",
+    "plt.legend()\n",
+    "\n",
+    "plt.show()"
    ]
   },
   {

learn-pr/azure/understand-classification-machine-learning/notebooks/6-5-exercise-assess-regression-model.ipynb

@@ -27,15 +27,20 @@
     "!pip install statsmodels\n",
     "!wget https://raw.githubusercontent.com/MicrosoftDocs/mslearn-introduction-to-machine-learning/main/graphing.py\n",
     "!wget https://raw.githubusercontent.com/MicrosoftDocs/mslearn-introduction-to-machine-learning/main/Data/avalanche.csv\n",
-    "import graphing # custom graphing code. See our GitHub repo for details\n",
+    "import matplotlib.pyplot as plt\n",
     "\n",
     "#Import the data from the .csv file\n",
     "dataset = pandas.read_csv('avalanche.csv', delimiter=\"\\t\")\n",
     "\n",
     "#Let's have a look at the data and the relationship we're going to model\n",
     "print(dataset.head())\n",
     "\n",
-    "graphing.box_and_whisker(dataset, label_x=\"avalanche\", label_y=\"weak_layers\")"
+    "plt.boxplot([dataset[\"avalanche\"], dataset[\"weak_layers\"]])\n",
+    "\n",
+    "plt.xlabel(\"Avalanches, Weak Layers\")\n",
+    "plt.ylabel(\"Values\")\n",
+    "\n",
+    "plt.show()"
    ]
   },
   {
@@ -133,10 +138,22 @@
    "metadata": {},
    "outputs": [],
    "source": [
+    "import matplotlib.pyplot as plt\n",
+    "import pandas as pd\n",
+    "\n",
     "def predict(weak_layers):\n",
-    "    return model.predict(dict(weak_layers=weak_layers))\n",
+    "    return model.predict(pd.DataFrame({\"weak_layers\": weak_layers}))\n",
+    "\n",
+    "plt.scatter(test['weak_layers'], test['avalanche'])\n",
+    "\n",
+    "trendline_values = predict(test['weak_layers'])\n",
+    "plt.plot(test['weak_layers'], trendline_values, color='red', label=\"Trendline\")\n",
+    "\n",
+    "plt.xlabel(\"weak_layers\")\n",
+    "plt.ylabel(\"avalanche\")\n",
+    "plt.legend()\n",
     "\n",
-    "graphing.scatter_2D(test, label_x=\"weak_layers\", label_y=\"avalanche\", trendline=predict)"
+    "plt.show()"
    ]
   },
   {
@@ -154,7 +171,20 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "graphing.scatter_2D(test, label_x=\"weak_layers\", label_y=\"avalanche\", x_range=[-20,20], trendline=predict)"
+    "\n",
+    "def predict(weak_layers):\n",
+    "    return model.predict(pd.DataFrame({\"weak_layers\": weak_layers}))\n",
+    "\n",
+    "\n",
+    "plt.scatter(test['weak_layers'], test['avalanche'])\n",
+    "trendline_values = predict(test['weak_layers'])\n",
+    "plt.plot(test['weak_layers'], trendline_values, color='red', label=\"Trendline\")\n",
+    "plt.xlim(-20, 20)\n",
+    "plt.xlabel(\"weak_layers\")\n",
+    "plt.ylabel(\"avalanche\")\n",
+    "plt.legend()\n",
+    "\n",
+    "plt.show()"
    ]
   },
   {

learn-pr/azure/understand-classification-machine-learning/notebooks/6-7-exercise-improve-classification.ipynb

@@ -27,7 +27,6 @@
     "!pip install statsmodels\n",
     "!wget https://raw.githubusercontent.com/MicrosoftDocs/mslearn-introduction-to-machine-learning/main/graphing.py\n",
     "!wget https://raw.githubusercontent.com/MicrosoftDocs/mslearn-introduction-to-machine-learning/main/Data/avalanche.csv\n",
-    "import graphing # custom graphing code. See our GitHub repo for details\n",
     "\n",
     "#Import the data from the .csv file\n",
     "dataset = pandas.read_csv('avalanche.csv', delimiter=\"\\t\", index_col=0)\n",
@@ -254,15 +253,44 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "graphing.model_to_surface_plot(model_with_interaction, [\"weak_layers\", \"wind\"], test)"
+    "import matplotlib.pyplot as plt\n",
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "from mpl_toolkits.mplot3d import Axes3D\n",
+    "\n",
+    "def predict(weak_layers, wind, surface_hoar=0, fresh_thickness=0, no_visitors=0):\n",
+    "    return model_with_interaction.predict(pd.DataFrame({\n",
+    "        \"weak_layers\": weak_layers,\n",
+    "        \"wind\": wind,\n",
+    "        \"surface_hoar\": surface_hoar,\n",
+    "        \"fresh_thickness\": fresh_thickness,\n",
+    "        \"no_visitors\": no_visitors\n",
+    "    }))\n",
+    "\n",
+    "# Generate a graph for weak_layers and wind\n",
+    "weak_layers = np.linspace(-20, 20, 100)\n",
+    "wind = np.linspace(-20, 20, 100)\n",
+    "weak_layers_grid, wind_grid = np.meshgrid(weak_layers, wind)\n",
+    "\n",
+    "predicted_values = predict(weak_layers_grid.ravel(), wind_grid.ravel()).to_numpy().reshape(weak_layers_grid.shape)\n",
+    "\n",
+    "fig = plt.figure()\n",
+    "ax = fig.add_subplot(111, projection='3d')\n",
+    "ax.plot_surface(weak_layers_grid, wind_grid, predicted_values, cmap='viridis')\n",
+    "\n",
+    "ax.set_xlabel(\"Weak Layers\")\n",
+    "ax.set_ylabel(\"Wind\")\n",
+    "ax.set_zlabel(\"Predicted Values\")\n",
+    "\n",
+    "plt.show()"
    ]
   },
   {
    "cell_type": "markdown",
    "id": "1d8af2ee",
    "metadata": {},
    "source": [
-    "The graph is interactive - rotate it and explore how there's a clear s-shaped relationship between the features and probability.\n",
+    "There's now a clear s-shaped relationship between the features and probability.\n",
     "\n",
     "Let's now look at the features that we've said can interact:"
    ]
@@ -274,7 +302,36 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "graphing.model_to_surface_plot(model_with_interaction, [\"no_visitors\", \"fresh_thickness\"], test)"
+    "import matplotlib.pyplot as plt\n",
+    "import numpy as np\n",
+    "import pandas as pd\n",
+    "from mpl_toolkits.mplot3d import Axes3D\n",
+    "\n",
+    "def predict(no_visitors, fresh_thickness, weak_layers=0, wind=0, surface_hoar=0):\n",
+    "    return model_with_interaction.predict(pd.DataFrame({\n",
+    "        \"no_visitors\": no_visitors,\n",
+    "        \"fresh_thickness\": fresh_thickness,\n",
+    "        \"weak_layers\": weak_layers,\n",
+    "        \"wind\": wind,\n",
+    "        \"surface_hoar\": surface_hoar\n",
+    "    }))\n",
+    "\n",
+    "# Generate the graph\n",
+    "no_visitors = np.linspace(-20, 20, 100)\n",
+    "fresh_thickness = np.linspace(-20, 20, 100)\n",
+    "no_visitors_grid, fresh_thickness_grid = np.meshgrid(no_visitors, fresh_thickness)\n",
+    "\n",
+    "predicted_values = predict(no_visitors_grid.ravel(), fresh_thickness_grid.ravel()).to_numpy().reshape(no_visitors_grid.shape)\n",
+    "\n",
+    "fig = plt.figure()\n",
+    "ax = fig.add_subplot(111, projection='3d')\n",
+    "ax.plot_surface(no_visitors_grid, fresh_thickness_grid, predicted_values, cmap='viridis')\n",
+    "\n",
+    "ax.set_xlabel(\"No Visitors\")\n",
+    "ax.set_ylabel(\"Fresh Thickness\")\n",
+    "ax.set_zlabel(\"Predicted Values\")\n",
+    "\n",
+    "plt.show()"
    ]
   },
   {