def interactionPlotPredicted(data, pr, featuresToExamine):
fig, axes, summary_df = info_plots.actual_plot_interact(
model=pr,
X=data,
features=featuresToExamine,
feature_names=featuresToExamine)
save("interactionPlotPredicted", plt=plt, fig=fig)
def actual_interact_plot(self, feature, var_name = None, sample = 10000, which_classes = None, show_outliers=True, **kargs):
fig, axes, result = info_plots.actual_plot_interact(
model = self.md, X = self.sample(sample),
features=feature, feature_names=var_name or feature,
which_classes=which_classes, show_outliers= show_outliers, **kargs)
self.actual_interact_data = ResultDF(result, 'count')
plt.show()
def pred_target_plot_inter(clf,X,Y,features,label,grid_ranges=None):
from pdpbox import info_plots
pd.options.mode.chained_assignment = None # Turn warning msg off
# Extract the classifier object from the clf multilearn object
index = Y.columns.to_list().index(label)
clf = clf.classifiers_[index]
clf.verbose = False #Turn verbose off after this to tidy prints
if(grid_ranges is None):
fig, ax, summary_df = info_plots.actual_plot_interact(clf,X,features=features,feature_names=features,grid_types=['equal','equal'])
else:
fig, ax, summary_df = info_plots.actual_plot_interact(clf,X,features=features,feature_names=features,grid_types=['equal','equal'],
show_outliers='True',grid_ranges=grid_ranges)
pd.options.mode.chained_assignment = 'warn' # Turn warning msg back on
clf.verbose = True # reset
return fig, ax
# %% target_plot_interact ----------------------------------------------------
fig, axes, summary_df = info_plots.target_plot_interact(
df=XY,
features=x_cols[2:],
feature_names=x_cols[2:],
target=y_cols[0],
)
# %% actual_plot_interact ----------------------------------------------------
fig, axes, summary_df = info_plots.actual_plot_interact(
model=model,
X=X,
features=x_cols[3:],
feature_names=x_cols[3:],
which_classes=[2, 5],
)
# %% pdp_interact: Preset ----------------------------------------------------
pdp_interacted_tmp = pdp.pdp_interact(
model=model,
dataset=X,
model_features=x_cols,
features=x_cols[:2],
num_grid_points=[10, 10],
percentile_ranges=[None, None],
n_jobs=1,
)
display(summary_df)
#And finally let's obtain the PDP for the feature Age
pdp_fare = pdp.pdp_isolate(
model=titanic_model, dataset=titanic_data, model_features=titanic_features, feature='Age'
)
fig, axes = pdp.pdp_plot(pdp_fare, 'Age', plot_pts_dist=True)
#Let's study the link between Age and Pclass
#Statistics of survivors based on Age and Pclass
fig, axes, summary_df = info_plots.target_plot_interact(
df=titanic_data, features=['Age', 'Pclass'], feature_names=['Age', 'Pclass'], target=titanic_target
)
display(summary_df.head())
#Prediction of our model, impact if Age and Pclass
fig, axes, summary_df = info_plots.actual_plot_interact(
model=titanic_model, X=titanic_data[titanic_features], features=['Age', 'Pclass'], feature_names=['Age', 'Pclass']
)
display(summary_df.head())
#PDP for the interaction between Age and Pclass
inter1 = pdp.pdp_interact(
model=titanic_model, dataset=titanic_data, model_features=titanic_features, features=['Age', 'Pclass']
)
fig, axes = pdp.pdp_interact_plot(
pdp_interact_out=inter1, feature_names=['age', 'Pclass'], plot_type='contour', x_quantile=True, plot_pdp=True
)
#Let's study the link between Fare and Sex
#Statistics of survivors based on Fare and Sex
fig, axes, summary_df = info_plots.target_plot_interact(
df=titanic_data, features=['Fare', 'Sex'], feature_names=['Fare', 'Sex'], target=titanic_target
)
feature_name=_,
target=target_variable
)
#PDP Plots : Actual Plot
for _ in X.columns.tolist():
fig,axes,summary_df=info_plots.actual_plot(
model=baseline,
X=df[X.columns],
feature=_,
feature_name=_,
predict_kwds={}
)
fig, axes, summary_df = info_plots.actual_plot_interact(
model=baseline, X=df[X.columns], features=interactions_2way, feature_names=interactions_2way
)
#PDP Plot : Grid Plot
interactions = pdp.pdp_interact(
model=baseline, dataset=df, model_features=X.columns, features=interactions_2way
)
fig, axes = pdp.pdp_interact_plot(interactions,interactions_2way , plot_type='grid', x_quantile=True, plot_pdp=False)
#SHAP
shap.initjs()
explainer=shap.TreeExplainer(baseline)
shap_values=explainer.shap_values(x_train[_feat])
shap.summary_plot(shap_values,x_train[_feat])
generateInsight(xgb_bin1,'longitude',x_train_bin1)
generateInsight(xgb_bin1,'accommodates',x_train_bin1)
generateInsight(xgb_bin1,'extra_people',x_train_bin1)
generateInsight(xgb_bin1,'minimum_nights',x_train_bin1)
generateInsight(xgb_bin1,'guests_included',x_train_bin1)
generateInsight(xgb_bin1,[col for col in x_train_bin1 if col.startswith('room_type')],x_train_bin1)
generateInsight(xgb_bin1,[col for col in x_train_bin1 if col.startswith('property_type')],x_train_bin1)
#interaction plot
feats = ['latitude', 'longitude']
p = pdp.pdp_interact(xgb_bin1, x_train_bin1, x_train_bin1.columns,feats)
pdp.pdp_interact_plot(p, feats)#the model is correctly able to find to classify the centre to london close to Buckingham palace as most expensive
fig, axes, summary_df = info_plots.actual_plot_interact(
model=xgb_bin1, X=x_train_bin1,
features=['accommodates', [col for col in x_train_bin1 if col.startswith('property_type')]],
feature_names=['accommodates', 'Property Type'],
)
#Plot Feature importances
plotFeatureImportances(xgb_bin1,x_train_bin1.columns)
##############Model for Price Bin 2#####################
y_train_bin2=y_train.loc[(y_train.price_bins==2),'price']
y_test_bin2=y_test.loc[y_test.price_bins==2,'price']
x_train_bin2=x_train.loc[y_train_bin2.index,:]
x_test_bin2=x_test.loc[y_test_bin2.index,:]
y_train_bin2=y_train_bin2.reset_index(drop=True)