Relevance Feedback Based Query Expansion Model Using Borda Count and Semantic Similarity Approach.

Area of Science:

• Information Retrieval
• Natural Language Processing
• Computer Science

Background:

• Pseudo-Relevance Feedback (PRF) is a standard technique for query expansion in information retrieval.
• Selecting optimal expansion terms is crucial for PRF effectiveness, as not all terms from retrieved documents are relevant.
• Existing individual term selection methods have limitations.

Purpose of the Study:

• To explore the potential of individual query expansion term selection methods.
• To combine multiple term selection methods to overcome individual weaknesses.
• To improve the overall performance of information retrieval systems through enhanced query expansion.

Main Methods:

• Investigated the performance of individual query expansion term selection methods.
• Employed the Borda count rank aggregation approach to combine multiple selection methods.
• Utilized a semantic similarity approach to refine term selection after rank aggregation.

Main Results:

• Experimental results show significant performance improvements compared to individual term selection methods.
• The proposed combined approach outperforms existing state-of-the-art methods.
• Demonstrated the effectiveness of integrating Borda count and semantic similarity for query expansion.

Conclusions:

• Combining multiple query expansion term selection methods enhances information retrieval performance.
• The Borda count and semantic similarity integration offers a robust strategy for PRF.
• This research provides a novel approach to optimize query expansion for improved search results.