一种融合本体和最小二乘支持向量机的主题爬行方法

摘要

There are many implementation approaches to focused crawler. However,it is difficult to maintain a high and stable crawling process. To solve this problem,a focused crawling approach based on ontology and LS-SVM(least squares support vector machine)was proposed. An LS-SVM classifier was created using theproblematic food ontology and applied to the classification of web pages. Experimental results show that the proposed approach can get a higher harvest rate than other focused crawling approaches.%针对现有的主题爬行方法存在收益率不高和不稳定的问题，融合本体和最小二乘支持向量机理论，提出一种主题爬行方法Ontology-LSSVM．该方法将本体作为抓取主题的背景知识表示，得到一组主题相关概念的集合，再将其在网页文本中出现的词频作为输入，构造一个主题相关度 LS-SVM 分类器，用于网页的分类．使用舆论热点的食品安全问题作为爬行主题领域，建立问题食品本体，进行实验．结果表明：在本实验条件下，本文方法相比基于 LS-SVM、基于本体和基于关键字的主题爬行，能够维持更高的收益率．