文本分类中多分类器的综合机制

文本分类是Internet文本信息处理的基础,该文通过对传统文本分类方法的研究,如支持向量机理论、多组判别分析、贝叶斯方法和中心向量法等分类方法,观察到分类器对于不同类别的文本其区分程度有所不同,因此提出了一种基于综合评价方法的多分类器决策机制。在参数训练过程中,引入了最优化理论中的直接搜索方法,形成一个容纳多个分类器的容器,它是各个分类器的最优化的组合,旨在获得最佳的分类精度。通过实验验证,得到了比较理想的分类效果。

Multi-classifier Classification Based on Integrative Evaluation

Text classification plays an important role in the internet information processing,so a multi-classifier based on integrative evaluation approach is presented in this paper.It can be observed that a text classifier has different precision on different text classes by investigating some approaches of text classification such as Support Vector Machine,Multi-Group Discrimination Analysis,Naive Bayes approach and Simple Vector Distance approach.Therefore,it is more effective approach that different text classifiers have to be integrated as a multi-classifier in order to improve the classification preclsion,and the integrative mechanism is derived from Integrative Evaluation.The classification container is an optimizing combination of difference classifiers,and the experiment results show that it can improve the precision of text classification.