With the development of the Internet, the information explosion problem comes into being and it is challenging for users to search for the information they needed from e-books. Although the book recommendation system can help users find their focuses, it is not applicable for visually impaired users when using ordinary visual reading methods for knowledge acquisition. Therefore, a book recommendation system that suits their behavior habits is required. In order to provide accurate and effective book sets for users, we propose an algorithm based on fusing their preferences. For intelligently ranking the candidate book sets and help users find the right book quickly, we propose a context-aware algorithm based on users’ attention. Meanwhile, we introduce an improved calculation method for users’ attention to solving the problem of inaccurate prediction on users’ current attention when their action history is cluttered. We use the self-attention to preserve the users’ reading tendencies during the reading process, analyze users’ personal features and book content features, and improve the accuracy of the recommendation by merging the feature space. Finally, the improved algorithm proposed and comparative experiments were employed on the dataset collecting from the China Blind Digital Library, and the effectiveness of the improvement is proved in each experimental comparison results.
Mendeley helps you to discover research relevant for your work.
CITATION STYLE
Yu, Z., Bu, J., Li, S., Wang, W., Tang, L., & Zhao, C. (2020). Research on book recommendation system for people with visual impairment based on fusion of preference and user attention. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 12376 LNCS, pp. 83–90). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-030-58796-3_11