摘要
为向网络用户提供多样化的搜索结果,本文通过相关文档和相关主题之间的映射刻画了相关文档空间,提出了主题多样性优化模型。该模型首先对相关排序结果聚类,然后计算主题基量和增量,最终以二者的组合作为排序依据。在采用AQUAINT语料库、TREC N51-N100查询课题和主题覆盖度评价指标的试验设定下验证了所提优化模型的有效性,并找到了最优的模型和参数配置。本文提出了利用相关排序结构特征的主题覆盖度优化模型,为网络搜索实践提供了翔实的参考数据。
To provide diversified search results with web users, the paper describes the document space based on the mappings between relevant documents and relevant topics; and introduces a model to optimize search result's topic coverage. The model first clusters relevant ranking list; calculates the basal and incremental parts of topics; and ranks search result according to the combination of the two. The model is validated through an experiment, in which AQUAINT collection, TREC N51-N100 topics and topic coverage are used and gives the optimized model and parameters for the optimizing model. It introduces a novel model for topic coverage with the structure of ranking list; and provides exhaustive data for web search practice.
出处
《情报学报》
CSSCI
北大核心
2016年第2期137-145,共9页
Journal of the China Society for Scientific and Technical Information