The Pseudo-relevance Feedback Model Based on Quantum Probability Theory

Relevance Model (RM) is one of typical and generally stable methods for the query expansion in information retrieval (IR). This paper presents a novel information retrieval model based on the quantum probability theory, and makes a preliminary exploration on the application of quantum model in pseudo-relevance feedback. In the document weight allocation framework, we propose two re-ranking approaches based on linear weight allocation and quantum interference weight allocation, respectively. The experimental results on standard TREC datasets show that the proposed model is effective and potential in information retrieval tasks. Keywordsinformation retrieval; pseudo relevance feedback; quantum probability theory; document weight allocation