ISART: A Generic Framework for Searching Books with Social Information

Area of Science:

• Computer Science
• Information Retrieval
• Digital Libraries

Background:

• Effective book search remains a challenge across diverse fields like e-commerce and digital humanities.
• While social information (ratings, tags, reviews) is abundant online, its integration into book search systems is underexplored.

Purpose of the Study:

• To develop a novel framework, Integrated Search And Recommendation Technology (IsArt), for leveraging rich social information to improve book search.
• To investigate the impact of user-generated content on the effectiveness of book retrieval systems.

Main Methods:

• Developed IsArt, a framework combining a content-based search engine with a Generalized Content-based Filtering model for social information.
• Employed a learning-to-rank technique to integrate diverse reranking signals from content, metadata, and social data.
• Utilized CLEF/INEX Social Book Search Evaluation datasets from 2011-2014 for experimentation.

Main Results:

• Demonstrated that integrating social information significantly enhances book search effectiveness.
• IsArt achieved state-of-the-art performance on the CLEF/INEX Social Book Search Evaluation datasets over four years.
• The proposed framework effectively combines content-based ranking with social information reranking.

Conclusions:

• Social information is a valuable resource for improving book search systems.
• The IsArt framework provides a robust and effective method for integrating diverse information sources for book discovery.
• This research advances the field of information retrieval by showcasing a practical application of social data in search technology.