IMPROVED WEBPAGE CLASSIFICATION TECHNOLOGY BASED ON FEEDFORWARD BACKPROPAGATION NEURAL NETWORK

With the increasing number of web pages in the Internet, the problem of automatically classifying web page directories has become more and more pressing. This paper proposes an improved webpage classification method based on weighted terminology of words to achieve automatic and efficient classification of webpages. The method uses a set of functions to classify a web document. The improved weighting technique based on terminology as proposed in this paper has obvious advantages for feature selection and feature extraction. It can greatly reduce the large size of the classifier and improve the efficiency of web page classification. The algorithm proposed in this paper was run and tested on a benchmark data set. Test results show that compared to most of the existing term weighting techniques this technology is more accurate and has other advantages.