schuetoliver
/
handson-ml
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Replace y_gen_faces with y_bad, and other little fixes (thanks Ian!)

main
Aurélien Geron 2021-03-02 15:09:30 +13:00
parent 3d418c0308
commit 5e9bec62bc
1 changed files with 18 additions and 9 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

27
09_unsupervised_learning.ipynb
View File

@ -1660,7 +1660,8 @@
"metadata": {},
"outputs": [],
"source": [
"log_reg.score(X_test, y_test)"
"log_reg_score = log_reg.score(X_test, y_test)\n",
"log_reg_score"
]
},
{
@ -1698,7 +1699,15 @@
"metadata": {},
"outputs": [],
"source": [
"pipeline.score(X_test, y_test)"
"pipeline_score = pipeline.score(X_test, y_test)\n",
"pipeline_score"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"How much did the error rate drop?"
]
},
{
@ -1707,14 +1716,14 @@
"metadata": {},
"outputs": [],
"source": [
"1 - (1 - 0.977777) / (1 - 0.968888)"
"1 - (1 - pipeline_score) / (1 - log_reg_score)"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"How about that? We reduced the error rate by over 28%! But we chose the number of clusters $k$ completely arbitrarily, we can surely do better. Since K-Means is just a preprocessing step in a classification pipeline, finding a good value for $k$ is much simpler than earlier: there's no need to perform silhouette analysis or minimize the inertia, the best value of $k$ is simply the one that results in the best classification performance."
"How about that? We reduced the error rate by over 35%! But we chose the number of clusters $k$ completely arbitrarily, we can surely do better. Since K-Means is just a preprocessing step in a classification pipeline, finding a good value for $k$ is much simpler than earlier: there's no need to perform silhouette analysis or minimize the inertia, the best value of $k$ is simply the one that results in the best classification performance."
]
},
{
@ -3473,11 +3482,12 @@
"outputs": [],
"source": [
"def plot_faces(faces, labels, n_cols=5):\n",
" faces = faces.reshape(-1, 64, 64)\n",
" n_rows = (len(faces) - 1) // n_cols + 1\n",
" plt.figure(figsize=(n_cols, n_rows * 1.1))\n",
" for index, (face, label) in enumerate(zip(faces, labels)):\n",
" plt.subplot(n_rows, n_cols, index + 1)\n",
" plt.imshow(face.reshape(64, 64), cmap=\"gray\")\n",
" plt.imshow(face, cmap=\"gray\")\n",
" plt.axis(\"off\")\n",
" plt.title(label)\n",
" plt.show()\n",
@ -3485,7 +3495,7 @@
"for cluster_id in np.unique(best_model.labels_):\n",
" print(\"Cluster\", cluster_id)\n",
" in_cluster = best_model.labels_==cluster_id\n",
" faces = X_train[in_cluster].reshape(-1, 64, 64)\n",
" faces = X_train[in_cluster]\n",
" labels = y_train[in_cluster]\n",
" plot_faces(faces, labels)"
]
@ -3709,7 +3719,6 @@
"n_darkened = 3\n",
"darkened = X_train[:n_darkened].copy()\n",
"darkened[:, 1:-1] *= 0.3\n",
"darkened = darkened.reshape(-1, 64*64)\n",
"y_darkened = y_train[:n_darkened]\n",
"\n",
"X_bad_faces = np.r_[rotated, flipped, darkened]\n",
@ -3819,7 +3828,7 @@
"metadata": {},
"outputs": [],
"source": [
"plot_faces(X_bad_faces, y_gen_faces)"
"plot_faces(X_bad_faces, y_bad)"
]
},
{
@ -3829,7 +3838,7 @@
"outputs": [],
"source": [
"X_bad_faces_reconstructed = pca.inverse_transform(X_bad_faces_pca)\n",
"plot_faces(X_bad_faces_reconstructed, y_gen_faces)"
"plot_faces(X_bad_faces_reconstructed, y_bad)"
]
},
{