An Improved Density-Based Method for Reducing Training Data in KNN

k-Nearest Neighbor (KNN) algorithm was an efficient text categorization algorithm in recall and accuracy, but the computational overhead of KNN was directly proportional to the sample size, so its classification speed was low in large-scale sample data. Aiming at this problem, the paper presented a density-based method for reducing training data, the method clustered each class of sample data into several clusters and reduced the noise sample data, and then combined some higher similar sample documents in each cluster into one document. Results of the experiment indicated that the method can reduce the computational overhead of KNN text classification, and the performance is approximately equal to those of the traditional KNN.