Research on Text Classification Algorithm of Largest Dispersion Based on Term Frequency

In order to achieve a document in accordance with the contents of the page automatic classification, put forward the largest dispersion of text classification algorithm based on the term frequency. The algorithm using backward term frequency algorithm for the n-types typical texts confirm the scientific and effective characteristics set of n-types; rely on it, getting the classification values of webpage documents in the n-types characteristics set through adopt to the largest dispersion algorithm, getting the largest dispersion after dispersion comparison; and then compared the largest dispersion value with relative threshold, if the value is larger than the threshold, it is the type of webpage documents, but if the value is smaller than the threshold, the judgement about the type of document is invalid. The algorithm has good robustness and easy-to-use, which is very effective for the large-scale data of small documents.