基于元样本稀疏表示分类器的文本资源分类

Metasample Based Sparse Representation Classification for Text Classifying

首先分析文本分类的现状,根据文本分类算法的要求和稀疏表示分类算法(SRC)的思想,设计基于元样本的稀疏表示分类器(MSRC),并应用于文本分类研究。实验结果表明,该MSRC算法具有较好的文本分类效果,有助于提高基于内容的信息检索效率。

Text classification is an important step in text preprocessing. Efficient text categorization can help to improve the efficiency of content-based information retrieval. This paper firstly analyzes the status of text classification study. Then, based on the requirements of text classification, the authors propose a metasample based sparse representation classification algorithm according to the theory of SRC algorithm. The experimental results on the text classification prove that the proposed algorithm is efficient.