The Design Of A Dynamic Text Classifier Based On VSM

This article introduces the design of a dynamic text classifier which can be applied to Chinese character input method program. The classifier is based on vector space model and categorize dynamic text stream which is typed by user from keyboard. The aim for categorization is to guide input program prompting repeated codes words.Firstly, this article introduces the background, content and purport of this research project; Secondly, summarizes correlate conception for text classifier, the process of constructing a text classifier and some technologies related with text categorization. Thirdly, gives the thoughtfulness of designing dynamic text classifier based on static text classifiers in existence, describes generating feature technologies, extracting feature technologies and the methods which are employed in constructing text classifier in particular, introduces the methods of simulating dynamic text. At last, simply introduces the implementation by coding for this dynamic text classifier.The dynamic text classifier based on VSM model introduced in this article adopts multi-classifier pattern, provides different class with different feature sub-space, adopts Rocchio method to construct classifier. Chinese character input program making use of this dynamic text classifier can reduce repeated code rate, or adjust the order of words with same code, and ultimately achieve the purpose of increasing the speed of inputting Chinese words. From this point, it makes sense to try to design such dynamic text classifier.