基于蚂蚁算法的Deep Web页面信息抽取方法研究

Study on Deep Web Information Extraction Technology Based on Ant Algorithm

针对煤炭监测数据的复杂多变性及Deep Web数据查询结果网页描述信息的特点,提出了一种基于蚂蚁算法和本体指导网页信息抽取的方法。首先构建基于简单本体的数据抽取系统,通过对结果页面中包含本体语义信息的数据的映像定位,结合蚂蚁算法分析信息素浓度在DOM树上的分布比较,实现数据块路径抽取规则算法及数据分割特征码的生成。以煤炭行业获取的数据进行抽取性能测试,数据实验表明,抽取算法结果具有较高的准确率。

Due to the complex of the coal monitoring data, a novel approach of web page information extraction guided by ant colony algorithm is proposed.The method first builded a simple ontology-based data extraction system. By positioning the image data on the result pages, and combined with ant algorithm, it creates extraction rules . The extraction performance test data obtained by the coal industry,and the experimental results indicate that the method gives a better accuracy according to the extraction.