基于改进SVM的文本混沌性分类优化技术实现  

Research on text chaos classification technology based on improved SVM

作　　者：赵颖[1] 

机构地区：[1]安顺学院,贵州安顺561000

出　　处：《现代电子技术》2016年第20期39-43,共5页Modern Electronics Technique

基　　金：国家自然科学基金(61304146)

摘　　要：当前的文本分类方法无法较好地处理海量文本以及文本特征空间数据,不能打破计算机处理性能和内存的约束,实现文本混沌性分类。而云计算平台可向用户提供需要的运算能力和存储空间。提出一种优化SVM的云计算环境下文本混沌性分类方法,设计Hadoop开源云计算系统,通过该系统中的Map Reduce模型对分类过程进行处理,提高分类的效率。采用优化SVM分类方法将混沌文本分类二次规划过程中的不等式限制变换成等式限制,提高海量文本混沌性分类精度。实验结果表明,所设计分类方法具有更高的处理效率,可以对海量文本数据进行准确的分类。The current text categorization methods are unable to deal with massive amounts of text and text feature space data better, and can＇t break the constraints of computer processing performance and memory and realize the chaotic text classification. The cloud computing platform can provide the computing capacity and storage space for users, so an optimized SVM based text chaos classification method effective in cloud computing environment is put forward. Hadoop open source cloud computing system was designed. The classification process is dealt with by Map Reduce model of the system to improve the classification efficiency. The optimized SVM classification method is used to convert the inequality constraints in the quadratic programming process of text chaos categorization into the equation constraints, so as to improve classification precision of massive chaotic texts. The experimental result shows that the classification method has higher treatment efficiency, and can classify the massive text data accurately.

关 键 词：云计算 文本分类 混沌分类 优化SVM 

分 类 号：TN911-34[电子电信—通信与信息系统] TP301[电子电信—信息与通信工程]