Improving Retrieval Effectiveness For Temporal-Constrained Top-K Query Processing

Large-scale Information Retrieval systems constantly need to strike a balance between effectiveness and efficiency. More effective methods often require longer query processing time. But if it takes too long to process a query, users would become dissatisfied and query load across servers might become unbalanced. Thus, it would be interesting to study how to process queries under temporal constraints so that search results for all queries can be returned within a specified time limit without significant effectiveness degradations. In this paper, we focus on top-K query processing for temporally constrained retrieval. The goal is to figure out what kind of query processing techniques should be used to meet the constraint on query processing time while minimizing the effectiveness loss of the search results. Specifically, we propose three temporal constrained top-K query processing techniques and then empirically evaluate them over TREC collections. Results show that all of the proposed techniques can meet the temporal constraints, and the document prioritization technique can return more effective search results.