Web文本分类中特征项权重的研究

The Study on Feature Items Weight of Web Text Classification

Web数据挖掘的主要研究内容是如何快速有效地在异构的、半结构化的、动态的Web信息资源获取有用信息,而Web文本分类是Web数据挖掘的核心。通过分析Web文本的特点,本文提出一种改进的TF-IWF特征项权重计算方法来对web文本进行分类。实验表明,该方法提高了web文本分类系统的分类性能,具有较好的正确率和召回率。

Web Data Mining major research how to get valuable information from heterogeneous, semi-structured, dynamic Web information resources quickly and efficiently, and Web text classification is the core of Web data mining. By analyzing the characteristics of Web text, this paper proposes a modified TF-IWF feature items weight calculation method. Experimental results show that, this method improves the Web text classification system performance and has good precision and recall rate.