revert to unaliased imports

Author: labarba

notebooks_en/4_Polynomial_Regression.ipynb

@@ -39,9 +39,9 @@
    "metadata": {},
    "outputs": [],
    "source": [
-    "from matplotlib import pyplot as plt\n",
+    "from matplotlib import pyplot\n",
     "from autograd import grad\n",
-    "import autograd.numpy as np"
+    "from autograd import numpy"
    ]
   },
   {
@@ -63,10 +63,10 @@
     }
    ],
    "source": [
-    "np.random.seed(0)    # fix seed for reproducibility\n",
-    "x = np.linspace(-3, 3, 20)\n",
-    "y = x**4 + x**3 - 4*x**2 + 8*np.random.normal(size=len(x))\n",
-    "plt.scatter(x, y);"
+    "numpy.random.seed(0)    # fix seed for reproducibility\n",
+    "x = numpy.linspace(-3, 3, 20)\n",
+    "y = x**4 + x**3 - 4*x**2 + 8*numpy.random.normal(size=len(x))\n",
+    "pyplot.scatter(x, y);"
    ]
   },
   {
@@ -111,7 +111,7 @@
     "def polynomial_features(x, degree):\n",
     "    \"\"\" Generate polynomial features for x.\"\"\"\n",
     "    \n",
-    "    X = np.empty((len(x), degree+1))\n",
+    "    X = numpy.empty((len(x), degree+1))\n",
     "    for i in range(degree+1):\n",
     "        X[:,i] = x**i\n",
     "    return X\n",
@@ -192,7 +192,7 @@
     "    Returns:\n",
     "      1D array of predicted values\n",
     "    '''\n",
-    "    return np.dot(X, params)\n",
+    "    return numpy.dot(X, params)\n",
     "\n",
     "def mse_loss(params, model, X, y):\n",
     "    '''\n",
@@ -206,7 +206,7 @@
     "      float, mean squared error\n",
     "    '''\n",
     "    y_pred = model(params, X)\n",
-    "    return np.mean( np.sum((y-y_pred)**2) )\n",
+    "    return numpy.mean( numpy.sum((y-y_pred)**2) )\n",
     "\n",
     "gradient = grad(mse_loss)"
    ]
@@ -265,13 +265,13 @@
    "source": [
     "max_iter = 3000\n",
     "alpha = 0.01\n",
-    "params = np.zeros(X_scaled.shape[1])\n",
-    "descent = np.ones(X_scaled.shape[1])\n",
+    "params = numpy.zeros(X_scaled.shape[1])\n",
+    "descent = numpy.ones(X_scaled.shape[1])\n",
     "i = 0\n",
     "\n",
     "from sklearn.metrics import mean_absolute_error\n",
     "\n",
-    "while np.linalg.norm(descent) > 0.01 and i < max_iter:\n",
+    "while numpy.linalg.norm(descent) > 0.01 and i < max_iter:\n",
     "    descent = gradient(params, linear_regression, X_scaled, y)\n",
     "    params = params - descent * alpha\n",
     "    loss = mse_loss(params, linear_regression, X_scaled, y)\n",
@@ -341,13 +341,13 @@
     }
    ],
    "source": [
-    "xgrid = np.linspace(x.min(), x.max(), 30)\n",
+    "xgrid = numpy.linspace(x.min(), x.max(), 30)\n",
     "Xgrid_poly_feat = polynomial_features(xgrid, degree)\n",
     "Xgrid_scaled = min_max_scaler.transform(Xgrid_poly_feat)\n",
     "Xgrid_scaled[:,0] = 1 \n",
-    "plt.scatter(x, y, c='r', label='true')\n",
-    "plt.plot(xgrid, Xgrid_scaled@params, label='predicted')\n",
-    "plt.legend();"
+    "pyplot.scatter(x, y, c='r', label='true')\n",
+    "pyplot.plot(xgrid, Xgrid_scaled@params, label='predicted')\n",
+    "pyplot.legend();"
    ]
   },
   {
@@ -489,7 +489,7 @@
     "      float, regularized mean squared error\n",
     "    '''\n",
     "    y_pred = model(params, X)\n",
-    "    return np.mean( np.sum((y-y_pred)**2) ) + _lambda * np.sum( params[1:]**2 )\n",
+    "    return numpy.mean( numpy.sum((y-y_pred)**2) ) + _lambda * numpy.sum( params[1:]**2 )\n",
     "\n",
     "gradient = grad(regularized_loss)                                                        "
    ]
@@ -529,13 +529,13 @@
    "source": [
     "max_iter = 3000\n",
     "alpha = 0.01\n",
-    "params = np.zeros(X_scaled.shape[1])\n",
-    "descent = np.ones(X_scaled.shape[1])\n",
+    "params = numpy.zeros(X_scaled.shape[1])\n",
+    "descent = numpy.ones(X_scaled.shape[1])\n",
     "i = 0\n",
     "\n",
     "from sklearn.metrics import mean_absolute_error\n",
     "\n",
-    "while np.linalg.norm(descent) > 0.01 and i < max_iter:\n",
+    "while numpy.linalg.norm(descent) > 0.01 and i < max_iter:\n",
     "    descent = gradient(params, linear_regression, X_scaled, y)\n",
     "    params = params - descent * alpha\n",
     "    loss = mse_loss(params, linear_regression, X_scaled, y)\n",
@@ -586,10 +586,10 @@
     "print(\"weights with regularization\")\n",
     "print(params)\n",
     "\n",
-    "plt.scatter(x, y, c='r', label='true')\n",
-    "plt.plot(xgrid, Xgrid_scaled@no_regularization_params, label='w/o regularization')\n",
-    "plt.plot(xgrid, Xgrid_scaled@params, label='with regularization')\n",
-    "plt.legend();"
+    "pyplot.scatter(x, y, c='r', label='true')\n",
+    "pyplot.plot(xgrid, Xgrid_scaled@no_regularization_params, label='w/o regularization')\n",
+    "pyplot.plot(xgrid, Xgrid_scaled@params, label='with regularization')\n",
+    "pyplot.legend();"
    ]
   },
   {
@@ -676,9 +676,9 @@
     "print(model.coef_)\n",
     "print(model.intercept_)\n",
     "\n",
-    "plt.scatter(x, y, c='r', label='true')\n",
-    "plt.plot(xgrid, y_pred_sklearn, label='sklearn ridge regression')\n",
-    "plt.legend();\n"
+    "pyplot.scatter(x, y, c='r', label='true')\n",
+    "pyplot.plot(xgrid, y_pred_sklearn, label='sklearn ridge regression')\n",
+    "pyplot.legend();\n"
    ]
   },
   {