After training the classifier, we can use the predict method, which takes in new data instances and returns the predicted corresponding labels (classes). It does so using averages of properties as calculated when fitting the estimator.

University of Michigan - Ann Arbor

- fit(X_train, y_train)
- predict(X_test)
- score(X_test, y_test)

Useful Methods for KNN Classifier

https://scikit-learn.org/stable/modules/generated/sklearn.neighbors.KNeighborsClassifier.html

Documentation for sklearn.neighbors.KNeighborsClassifier

The score method returns a float between 0 and 1 indicating the percentage of test labels predicted correctly. This can be understood as calling something like predict() on X_test and then comparing the output labels to y_test. A high percentage indicates greater accuracy. This method can be used after the model has been fitted

score(X_test, y_test) for sklearn.neighbors.KNeighborsClassifier

predict(X_test) for sklearn.neighbors.KNeighborsClassifier

After splitting the data and creating the knn classifier object, we can use the fit method, which trains ("fits") the model using the training data. This is necessary so that the classifier can predict the labels of test data. Note that X_train is a dataframe of items and their features, whereas y_train is a series of corresponding true labels.

Learn Before

Related