A Fast Document Classification Algorithm Based on Improved KNN

A novel KNN classification algorithm combining model and evidence theory is proposed in this paper. The new method not only overcomes the main shortage of lazy learning in traditional KNN, but also takes the distances between samples to be recognized and samples in k-neighbors into account. At the same time the method resolves the unrecognizable cases of unknown samples. Applying the classification algorithm into the document recognition, experimental results show its satisfied recognition rate and fast categorization speed