Pushing the docs to dev/ for branch: master, commit 83beb5f35fe88b4d56d8cf9eb95f9617e456f6a4

Author: sklearn-ci

dev/_downloads/auto_examples_jupyter.zip

@@ -24,7 +24,7 @@
       "execution_count": null, 
       "cell_type": "code", 
       "source": [
-        "print(__doc__)\n\n# Authors: Jan Hendrik Metzen <[email protected]>\n# License: BSD 3 clause\n\n\nimport time\n\nimport numpy as np\n\nimport matplotlib.pyplot as plt\n\nfrom sklearn.kernel_ridge import KernelRidge\nfrom sklearn.model_selection import GridSearchCV\nfrom sklearn.gaussian_process import GaussianProcessRegressor\nfrom sklearn.gaussian_process.kernels import WhiteKernel, ExpSineSquared\n\nrng = np.random.RandomState(0)\n\n# Generate sample data\nX = 15 * rng.rand(100, 1)\ny = np.sin(X).ravel()\ny += 3 * (0.5 - rng.rand(X.shape[0]))  # add noise\n\n# Fit KernelRidge with parameter selection based on 5-fold cross validation\nparam_grid = {\"alpha\": [1e0, 1e-1, 1e-2, 1e-3],\n              \"kernel\": [ExpSineSquared(l, p)\n                         for l in np.logspace(-2, 2, 10)\n                         for p in np.logspace(0, 2, 10)]}\nkr = GridSearchCV(KernelRidge(), cv=5, param_grid=param_grid)\nstime = time.time()\nkr.fit(X, y)\nprint(\"Time for KRR fitting: %.3f\" % (time.time() - stime))\n\ngp_kernel = ExpSineSquared(1.0, 5.0, periodicity_bounds=(1e-2, 1e1)) \\\n    + WhiteKernel(1e-1)\ngpr = GaussianProcessRegressor(kernel=gp_kernel)\nstime = time.time()\ngpr.fit(X, y)\nprint(\"Time for GPR fitting: %.3f\" % (time.time() - stime))\n\n# Predict using kernel ridge\nX_plot = np.linspace(0, 20, 10000)[:, None]\nstime = time.time()\ny_kr = kr.predict(X_plot)\nprint(\"Time for KRR prediction: %.3f\" % (time.time() - stime))\n\n# Predict using kernel ridge\nstime = time.time()\ny_gpr = gpr.predict(X_plot, return_std=False)\nprint(\"Time for GPR prediction: %.3f\" % (time.time() - stime))\n\nstime = time.time()\ny_gpr, y_std = gpr.predict(X_plot, return_std=True)\nprint(\"Time for GPR prediction with standard-deviation: %.3f\"\n      % (time.time() - stime))\n\n# Plot results\nplt.figure(figsize=(10, 5))\nlw = 2\nplt.scatter(X, y, c='k', label='data')\nplt.plot(X_plot, np.sin(X_plot), color='navy', lw=lw, label='True')\nplt.plot(X_plot, y_kr, color='turquoise', lw=lw,\n         label='KRR (%s)' % kr.best_params_)\nplt.plot(X_plot, y_gpr, color='darkorange', lw=lw,\n         label='GPR (%s)' % gpr.kernel_)\nplt.fill_between(X_plot[:, 0], y_gpr - y_std, y_gpr + y_std, color='darkorange',\n                 alpha=0.2)\nplt.xlabel('data')\nplt.ylabel('target')\nplt.xlim(0, 20)\nplt.ylim(-4, 4)\nplt.title('GPR versus Kernel Ridge')\nplt.legend(loc=\"best\",  scatterpoints=1, prop={'size': 8})\nplt.show()"
+        "print(__doc__)\n\n# Authors: Jan Hendrik Metzen <[email protected]>\n# License: BSD 3 clause\n\n\nimport time\n\nimport numpy as np\n\nimport matplotlib.pyplot as plt\n\nfrom sklearn.kernel_ridge import KernelRidge\nfrom sklearn.model_selection import GridSearchCV\nfrom sklearn.gaussian_process import GaussianProcessRegressor\nfrom sklearn.gaussian_process.kernels import WhiteKernel, ExpSineSquared\n\nrng = np.random.RandomState(0)\n\n# Generate sample data\nX = 15 * rng.rand(100, 1)\ny = np.sin(X).ravel()\ny += 3 * (0.5 - rng.rand(X.shape[0]))  # add noise\n\n# Fit KernelRidge with parameter selection based on 5-fold cross validation\nparam_grid = {\"alpha\": [1e0, 1e-1, 1e-2, 1e-3],\n              \"kernel\": [ExpSineSquared(l, p)\n                         for l in np.logspace(-2, 2, 10)\n                         for p in np.logspace(0, 2, 10)]}\nkr = GridSearchCV(KernelRidge(), cv=5, param_grid=param_grid)\nstime = time.time()\nkr.fit(X, y)\nprint(\"Time for KRR fitting: %.3f\" % (time.time() - stime))\n\ngp_kernel = ExpSineSquared(1.0, 5.0, periodicity_bounds=(1e-2, 1e1)) \\\n    + WhiteKernel(1e-1)\ngpr = GaussianProcessRegressor(kernel=gp_kernel)\nstime = time.time()\ngpr.fit(X, y)\nprint(\"Time for GPR fitting: %.3f\" % (time.time() - stime))\n\n# Predict using kernel ridge\nX_plot = np.linspace(0, 20, 10000)[:, None]\nstime = time.time()\ny_kr = kr.predict(X_plot)\nprint(\"Time for KRR prediction: %.3f\" % (time.time() - stime))\n\n# Predict using gaussian process regressor\nstime = time.time()\ny_gpr = gpr.predict(X_plot, return_std=False)\nprint(\"Time for GPR prediction: %.3f\" % (time.time() - stime))\n\nstime = time.time()\ny_gpr, y_std = gpr.predict(X_plot, return_std=True)\nprint(\"Time for GPR prediction with standard-deviation: %.3f\"\n      % (time.time() - stime))\n\n# Plot results\nplt.figure(figsize=(10, 5))\nlw = 2\nplt.scatter(X, y, c='k', label='data')\nplt.plot(X_plot, np.sin(X_plot), color='navy', lw=lw, label='True')\nplt.plot(X_plot, y_kr, color='turquoise', lw=lw,\n         label='KRR (%s)' % kr.best_params_)\nplt.plot(X_plot, y_gpr, color='darkorange', lw=lw,\n         label='GPR (%s)' % gpr.kernel_)\nplt.fill_between(X_plot[:, 0], y_gpr - y_std, y_gpr + y_std, color='darkorange',\n                 alpha=0.2)\nplt.xlabel('data')\nplt.ylabel('target')\nplt.xlim(0, 20)\nplt.ylim(-4, 4)\nplt.title('GPR versus Kernel Ridge')\nplt.legend(loc=\"best\",  scatterpoints=1, prop={'size': 8})\nplt.show()"
       ], 
       "outputs": [], 
       "metadata": {

dev/_downloads/auto_examples_python.zip

@@ -91,7 +91,7 @@
 y_kr = kr.predict(X_plot)
 print("Time for KRR prediction: %.3f" % (time.time() - stime))
 
-# Predict using kernel ridge
+# Predict using gaussian process regressor
 stime = time.time()
 y_gpr = gpr.predict(X_plot, return_std=False)
 print("Time for GPR prediction: %.3f" % (time.time() - stime))