Density Clustering Algorithm Based On Improved Support Vector Machine

The amount of information is increasing in a geometric progression in the 21st century. But as individuals, searching the information we needed quickly and effectively is very difficult. Under the urgent need, the technology of text automatic classification emerges. This technology classifies the information based on text document, and according to the model whih has built, the text classification processes were done by computers automatically.There are many problems in the process of text classification, such as the text vector of large sparse, high dimensions and great correlations among features, SVM provided solutions to solve these problems with clear aim. Therefore, SVM was strongly advocated in the text classification applications.As stated in dialectics: All things are the overall unity of opposites. Using SVM in text classification applications will not be easy to introduce. With the expansion and in-depth of applications step by step, a great number of new issues which haven't work out have arisen on SVM oen by one. This article is from the Chinese version of the process of text classification to reduce the number of vectors to improve efficiency and accuracy rate of classification point of view, to speed up the training of support vector machine classifier speed. Proposed an improved clustering method based on density of training samples extracted data play a decisive role to focus on the classification of support vector set as a new set of training samples to carry out classifier training.