面向Web信息检索的虚核文本分类算法

提出一种虚核文本分类算法。通过单类别下标记数据的特征词频计算该类别虚核在每个特征项处的特征引力场强，进而获得类别虚核，根据类别虚核对待标记文本产生的引力大小判断其所属类别。实验结果表明，与k近邻算法和朴素贝叶斯算法相比，虚核算法在分类精度和时间开销方面具有较大的优势。

Virtual Kernel Text Categorization Algorithm for Web Information Retrieval

This paper proposes a novel Virtual Kernel(VK) text categorization algorithm.It obtains a virtual kernel for each categorize through evaluating the contribution of each feature term.When an unlabeled text comes,the algorithm can give the categorize it belongs to through evaluating the attraction which is from every virtual kernel.Experimental results show that compared with the k Nearest Neighbor(kNN) and na ve Bayes algorithms,the proposed algorithm is remarkably superior to them in terms of classification efficiency and time cost.