Please use this identifier to cite or link to this item:

Title: An Effective Pareto Optimality Based Fusion Technique for Information Retrieval
Authors: Batri, Krishnan
Keywords: Information Retrieval;Data Fusion;Meta Search;Vector Space Model;Similarity Measures;Extended Boolean Model
Date: 2013-09
Issue Date: 2016-08-16 16:11:01 (UTC+8)
Abstract: Information Retrieval (IR) is the process of retrieving information that is relevant to the users' needs. Over the years, researchers tend to develop the best retrieval strategy, which achieves the best possible performance across all document collections. Their results indicate a pattern of tug-of-war relationship prevalent among the existing strategies, where in one strategy dominates the remaining strategies over other document collections. Data Fusion may nullify the aforesaid tug-of-war effect. It can extract the best possible performance among the participating members. Data Fusion in IR usually combines the various retrieval schemes (strategies) to enhance the overall system performance. Our proposed fusion functions assign relevance scores by considering non dependency among all participating strategies. Relevance score assignment based on the relationship between that specific document and all other documents in the corpus. The existing Comb functions treated as the baseline functions for our proposed functions. Proposed and baseline functions' performance tested among three medium size corpuses. The average precision value of functions indicates that, one of our proposed functions achieves better performance in comparison with the base line functions. The statistical analysis confirms the same.
Relation: 資管評論, 19(1), 61-80
MIS review
Data Type: article
DOI 連結:
Appears in Collections:[資管評論] 期刊論文

Files in This Item:

File SizeFormat
19(1)-61-80.pdf1116KbAdobe PDF504View/Open

All items in 學術集成 are protected by copyright, with all rights reserved.

社群 sharing