\n",
"
\n",
" Visualization omitted, Javascript library not loaded!
\n",
" Have you run `initjs()` in this notebook? If this notebook was from another\n",
" user you must also trust this notebook (File -> Trust notebook). If you are viewing\n",
" this notebook on github the Javascript has been stripped for security. If you are using\n",
" JupyterLab this error is because a JupyterLab extension has not yet been written.\n",
"
\n",
" Visualization omitted, Javascript library not loaded!
\n",
" Have you run `initjs()` in this notebook? If this notebook was from another\n",
" user you must also trust this notebook (File -> Trust notebook). If you are viewing\n",
" this notebook on github the Javascript has been stripped for security. If you are using\n",
" JupyterLab this error is because a JupyterLab extension has not yet been written.\n",
"
\n",
" "
],
"text/plain": [
""
]
},
"execution_count": 4,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"# visualize all the predictions' explanations for the 'setosa' class\n",
"shap.plots.force(shap_values[..., 0])"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Support vector machine with a linear kernel"
]
},
{
"cell_type": "code",
"execution_count": 5,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Accuracy = 100.0%\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"Using 120 background data samples could cause slower run times. Consider using shap.sample(data, K) or shap.kmeans(data, K) to summarize the background as K samples.\n"
]
},
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "9ea8f8a814e44baba56bd6649eeea0f6",
"version_major": 2,
"version_minor": 0
},
"text/plain": [
" 0%| | 0/30 [00:00\n",
"\n",
" Visualization omitted, Javascript library not loaded!
\n",
" Have you run `initjs()` in this notebook? If this notebook was from another\n",
" user you must also trust this notebook (File -> Trust notebook). If you are viewing\n",
" this notebook on github the Javascript has been stripped for security. If you are using\n",
" JupyterLab this error is because a JupyterLab extension has not yet been written.\n",
"
\n",
" "
],
"text/plain": [
""
]
},
"execution_count": 5,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"svc_linear = sklearn.svm.SVC(kernel=\"linear\", probability=True)\n",
"svc_linear.fit(X_train, Y_train)\n",
"print_accuracy(svc_linear.predict)\n",
"\n",
"# explain all the predictions in the test set\n",
"explainer = shap.Explainer(svc_linear.predict_proba, X_train)\n",
"shap_values = explainer(X_test)\n",
"\n",
"# visualize the training set predictions\n",
"shap.plots.force(shap_values[..., 0])"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Support vector machine with a radial basis function kernel"
]
},
{
"cell_type": "code",
"execution_count": 6,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Accuracy = 100.0%\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"Using 120 background data samples could cause slower run times. Consider using shap.sample(data, K) or shap.kmeans(data, K) to summarize the background as K samples.\n"
]
},
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "d77859c2b23f410c9d0aea1b5f474516",
"version_major": 2,
"version_minor": 0
},
"text/plain": [
" 0%| | 0/30 [00:00\n",
"\n",
" Visualization omitted, Javascript library not loaded!
\n",
" Have you run `initjs()` in this notebook? If this notebook was from another\n",
" user you must also trust this notebook (File -> Trust notebook). If you are viewing\n",
" this notebook on github the Javascript has been stripped for security. If you are using\n",
" JupyterLab this error is because a JupyterLab extension has not yet been written.\n",
"
\n",
" "
],
"text/plain": [
""
]
},
"execution_count": 6,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"svc_rbf = sklearn.svm.SVC(kernel=\"rbf\", probability=True)\n",
"svc_rbf.fit(X_train, Y_train)\n",
"print_accuracy(svc_rbf.predict)\n",
"\n",
"# explain all the predictions in the test set\n",
"explainer = shap.Explainer(svc_rbf.predict_proba, X_train)\n",
"shap_values = explainer(X_test)\n",
"\n",
"# visualize the training set predictions\n",
"shap.plots.force(shap_values[..., 0])"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Logistic regression"
]
},
{
"cell_type": "code",
"execution_count": 7,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Accuracy = 100.0%\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"Using 120 background data samples could cause slower run times. Consider using shap.sample(data, K) or shap.kmeans(data, K) to summarize the background as K samples.\n"
]
},
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "3a63d0290d4744c6b57067acba3f9d77",
"version_major": 2,
"version_minor": 0
},
"text/plain": [
" 0%| | 0/30 [00:00\n",
"\n",
" Visualization omitted, Javascript library not loaded!
\n",
" Have you run `initjs()` in this notebook? If this notebook was from another\n",
" user you must also trust this notebook (File -> Trust notebook). If you are viewing\n",
" this notebook on github the Javascript has been stripped for security. If you are using\n",
" JupyterLab this error is because a JupyterLab extension has not yet been written.\n",
"
\n",
" "
],
"text/plain": [
""
]
},
"execution_count": 7,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"linear_lr = sklearn.linear_model.LogisticRegression(solver=\"newton-cg\")\n",
"linear_lr.fit(X_train, Y_train)\n",
"print_accuracy(linear_lr.predict)\n",
"\n",
"# explain all the predictions in the test set\n",
"explainer = shap.Explainer(linear_lr.predict_proba, X_train)\n",
"shap_values = explainer(X_test)\n",
"\n",
"# visualize the training set predictions\n",
"shap.plots.force(shap_values[..., 0])"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Decision tree"
]
},
{
"cell_type": "code",
"execution_count": 8,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Accuracy = 100.0%\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"Using 120 background data samples could cause slower run times. Consider using shap.sample(data, K) or shap.kmeans(data, K) to summarize the background as K samples.\n"
]
},
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "70699397a390482f8a42a149bdb60e1d",
"version_major": 2,
"version_minor": 0
},
"text/plain": [
" 0%| | 0/30 [00:00\n",
"\n",
" Visualization omitted, Javascript library not loaded!
\n",
" Have you run `initjs()` in this notebook? If this notebook was from another\n",
" user you must also trust this notebook (File -> Trust notebook). If you are viewing\n",
" this notebook on github the Javascript has been stripped for security. If you are using\n",
" JupyterLab this error is because a JupyterLab extension has not yet been written.\n",
"
\n",
" "
],
"text/plain": [
""
]
},
"execution_count": 8,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"import sklearn.tree\n",
"\n",
"dtree = sklearn.tree.DecisionTreeClassifier(min_samples_split=2)\n",
"dtree.fit(X_train, Y_train)\n",
"print_accuracy(dtree.predict)\n",
"\n",
"# explain all the predictions in the test set\n",
"explainer = shap.Explainer(dtree.predict_proba, X_train)\n",
"shap_values = explainer(X_test)\n",
"\n",
"# visualize the training set predictions\n",
"shap.plots.force(shap_values[..., 0])"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Random forest"
]
},
{
"cell_type": "code",
"execution_count": 9,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Accuracy = 100.0%\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"Using 120 background data samples could cause slower run times. Consider using shap.sample(data, K) or shap.kmeans(data, K) to summarize the background as K samples.\n"
]
},
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "21620db188a64c9185448d836fa48e45",
"version_major": 2,
"version_minor": 0
},
"text/plain": [
" 0%| | 0/30 [00:00\n",
"\n",
" Visualization omitted, Javascript library not loaded!
\n",
" Have you run `initjs()` in this notebook? If this notebook was from another\n",
" user you must also trust this notebook (File -> Trust notebook). If you are viewing\n",
" this notebook on github the Javascript has been stripped for security. If you are using\n",
" JupyterLab this error is because a JupyterLab extension has not yet been written.\n",
"
\n",
" "
],
"text/plain": [
""
]
},
"execution_count": 9,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"from sklearn.ensemble import RandomForestClassifier\n",
"\n",
"rforest = RandomForestClassifier(n_estimators=100, max_depth=None, min_samples_split=2, random_state=0)\n",
"rforest.fit(X_train, Y_train)\n",
"print_accuracy(rforest.predict)\n",
"\n",
"# explain all the predictions in the test set\n",
"explainer = shap.Explainer(rforest.predict_proba, X_train)\n",
"shap_values = explainer(X_test)\n",
"\n",
"# visualize the training set predictions\n",
"shap.plots.force(shap_values[..., 0])"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Neural network"
]
},
{
"cell_type": "code",
"execution_count": 10,
"metadata": {},
"outputs": [
{
"name": "stdout",
"output_type": "stream",
"text": [
"Accuracy = 96.66666666666667%\n"
]
},
{
"name": "stderr",
"output_type": "stream",
"text": [
"Using 120 background data samples could cause slower run times. Consider using shap.sample(data, K) or shap.kmeans(data, K) to summarize the background as K samples.\n"
]
},
{
"data": {
"application/vnd.jupyter.widget-view+json": {
"model_id": "893fe22dc360430980bab9d2e5abac4f",
"version_major": 2,
"version_minor": 0
},
"text/plain": [
" 0%| | 0/30 [00:00\n",
"\n",
" Visualization omitted, Javascript library not loaded!
\n",
" Have you run `initjs()` in this notebook? If this notebook was from another\n",
" user you must also trust this notebook (File -> Trust notebook). If you are viewing\n",
" this notebook on github the Javascript has been stripped for security. If you are using\n",
" JupyterLab this error is because a JupyterLab extension has not yet been written.\n",
"
\n",
" "
],
"text/plain": [
""
]
},
"execution_count": 10,
"metadata": {},
"output_type": "execute_result"
}
],
"source": [
"from sklearn.neural_network import MLPClassifier\n",
"\n",
"nn = MLPClassifier(solver=\"lbfgs\", alpha=1e-1, hidden_layer_sizes=(5, 2), random_state=0)\n",
"nn.fit(X_train, Y_train)\n",
"print_accuracy(nn.predict)\n",
"\n",
"# explain all the predictions in the test set\n",
"explainer = shap.Explainer(nn.predict_proba, X_train)\n",
"shap_values = explainer(X_test)\n",
"\n",
"# visualize the training set predictions\n",
"shap.plots.force(shap_values[..., 0])"
]
},
{
"cell_type": "markdown",
"metadata": {},
"source": [
"## Summary Plot\n",
"\n",
"While force plots are great for seeing how each feature contributes to a single prediction, a beeswarm plot can show the relative importance of features across the entire test set. Here we show the beeswarm plot for the 'setosa' species using the Random Forest model."
]
},
{
"cell_type": "code",
"execution_count": null,
"metadata": {},
"outputs": [],
"source": [
"explainer = shap.Explainer(rforest.predict_proba, X_train)\n",
"shap_values = explainer(X_test)\n",
"\n",
"# summarize the effects of all the features\n",
"shap.plots.beeswarm(shap_values[..., 0])"
]
}
],
"metadata": {
"anaconda-cloud": {},
"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",
"version": "3.11.6"
}
},
"nbformat": 4,
"nbformat_minor": 1
}
