Hands-On Machine Learning with C++

kNN method

The kNN is a popular classification method that is sometimes used in regression problems. It is one of the most natural approaches to classification. The essence of the method is to classify the current item by the most prevailing class of its neighbors. Formally, the basis of the method is the hypothesis of compactness: if the metric of the distance between the examples is clarified successfully, then similar examples are more likely to be in the same class. For example, if you don't know what type of product to specify in the ad for a Bluetooth headset, you can find five similar headset ads. If four of them are categorized as Accessories and only one as Hardware, common sense will tell you that your ad should probably be in the ...