联合贝叶斯推理与遗传算法的主题信息搜索策略

A Searching Strategy of Topic Search Based on Bayesian Reasoning and Genetic Algorithm

为提高大数据环境下主题信息搜索的准确率和查全率,提出了将贝叶斯推理与遗传算法相结合的搜索策略.利用贝叶斯推理对文档的主题相关度进行了计算,并结合遗传算法对搜索过程进行启发式引导,同时引入差异度参数,在Heritrix框架基础上,利用集成开发环境Eclipse 3.3实现了相应功能.实验结果表明:搜索策略改进后的系统抓取主题页面所占比例与原系统相比有较大提高.

To improve the precision rate and recall rate of topic search in big data environment, this paper proposes a searching strategy based on Bayesian reasoning and genetic algorithm. It distinguishes the correlation between the web pages by Bayesian reasoning, and by genetic algorithm inspired pilot the searching process of and introduces the parameter of differentia. Based on Heritrix, the functions were implemented in the integrated development environment of Eclipse 3.3. The experimental results show that the new strategy improves the proportion between the topic page number and the total number of pages.