基于半监督学习的查询扩展模型

摘要

Query expansion is a optimization method for "word mismatch" issues in information retrieval domain. By analyzing the shortcomings of existing methods,query expansion model based on semi-supervised learning is proposed,the model seems query expansion as a classification problem,and using transductvie support vector machine to train the samples. Experiments show that the recall and precision rates of search engine are further improved by this method.%查询扩展是针对信息检索中常见的“词不匹配”问题提出的一种优化方法.通过分析现有查询扩展方法的不足,提出一种基于半监督学习的查询扩展模型,该模型将查询扩展看作一个分类问题,并采用直推式支持向量机对样本进行训练.实验结果表明该方法进一步提高了搜索引擎的查全率和查准率.