Author(s): |
Yong Li, Department of Computer Science and Technology, Binzhou University, Binzhou Shandong,China Mingyu Lu, Institute of Information Science and Technology, Dalian Maritime University, Dalian Liaoning, China Xinling Gan, CVIC Softeware, Middleware, Jinan, China |
Abstract: |
Basing on the characteristics of EBM web page, a method of classification to webpage was pre- sented, applying the Weighted Naive Bayesian (WNB). Firstly, the critical information was extracted by LUHN method, and then word weight was adjusted by evaluation functions, of which the WNB classifier was constructed. Lastly, the WNB classification was promoted. It is confirmed by our experiments that this method can reduce the data range and it can increase 6-9 percentage points than the traditional NB calcula- tion.
|