基于局部类别分析和遗传算法的查询优化算法

QUERY OPTIMIZATION ALGORITHM BASED ON LOCAL CLASSIFICATION ANALYSIS AND GENETIC ALGORITHM

查询扩展是信息检索中优化查询的一种有效方法。针对信息检索中用户查询关键词与文档标引词不匹配的问题,提出一种基于局部类别分析和遗传算法的查询优化算法。该算法分两个阶段实现:第1阶段对用户提交的查询Qold进行扩展,采用基于局部类别分析的查询扩展方法选择查询扩展词构成新查询Qnew;第2阶段对新查询Qnew进行权重分配,采用遗传算法对扩展后的查询进行权重调整得到最优查询向量,再次对测试集中的文档进行二次检索。实验结果表明,该算法比单独使用局部上下文分析算法、局部类别分析算法均有更优的检索性能。

Query expansion is an effective method for optimization query in information retrieval. To tackle the problem of mismatch between user query keywords and document tagging words during information retrieval, the authors put forward a query optimization algorithm based on local classification analysis and genetic algorithm. Th algorithm is realized in two stages. In the first stage, the user submitted query, Q,,1,z, is expanded, using a query expansion method based on local classification analysis to choose a query expansion word to build a new query, Q,eu.; in the second state, weight allocation is carried out for the new query, Q using genetic algorithm to adjust weight for the expanded query to obtain the optimal query vector, then carry out quadratic search for documents in the test suite. Experiment results show that the algorithm is superior at retrieval performance to either local context analysis algorithm alone or local classification analysis algorithm alone.