The Improved Naive Bayesian WEB Text Classification Algorithm
Ping Bai, Junqing Li · 2009
It is a very important task that how to classify Web pages automatically and effectively in accordance with the given model for machine learning. The traditional operation modes, including artificial way and semiautomatic way, form category abstracts after domain experts' personnel inspection and then put the results into a particular class library according to the scheduled requirements. An improved naive Bayesian Web text classification algorithm is proposed in this paper. The common Bayesian classifier assumes that all the items are equally important while in this paper the terms in each title are considered to be more important than others. Experiments showed that, the improved naive Bayesian algorithm is more precise in the text classification.