具有分类数据的 sklearn 树的混淆矩阵

Question

I have trained a sklearn tree with default parameters on a dataset in which all data are categorical. The dataset is about cars.

The rightmost column is the label associated with the attributes. Each row is an input instance. Because sklearn cannot handle categorical data, I used a dictionary to map each possible input for a feature into a dictionary and mapped them to positive integers so the dataset is no longer categorical and now consists of only integers where each has a meaning for each feature. I used the new training set to train the classifier.

#x_train= training dat
#y_train= training data label
from sklearn.tree import DecisionTreeClassifier
clf=DecisionTreeClassifier()
clf.fit(x_train,y_train)

I'm able to calculate accuracy for an arbitrary x_test and y_test input with score(x_test, y_test) . The question is how can I draw and visualize the confusion matrix for x_test and y_test ?

Thanks.

Answer 1

You can use plot_confusion_matrix to visually represent a confusion matrix. You only need to pass the fitted estimator ( clf in your case) along with the input ( X_test ) and the true target values ( y_test ). Here is an example:

import matplotlib.pyplot as plt
from sklearn.tree import DecisionTreeClassifier


clf = DecisionTreeClassifier()
clf.fit(X_train, y_train)

plot_confusion_matrix(clf, X_test, y_test)
plt.show()

The plot can further be customized eg by providing labels other than the ones in y_test . For more information, have a look at the documentation .