Fuzzy aggregation for results merging in information retrieval: An application of Choquet Integral

摘要

Peer-to-Peer Information Retrieval System (P2PIR) is a system that can query multiple information retrieval systems and merges ranked results list into a single result of documents. Classical methods are generally based on linear combination schemes. A major shortcoming of the classical methods is that there is no defined way to study dependencies and interactions existing among relevance criteria. In this paper, we propose a result merging method based on the Choquet Integral, called Choquet-Based Merging (CBM). The experimental results obtained on the test collection provided by TREC Contextual Suggestion track shows the effectiveness of our proposal.