SemBRS: A Semantic Analysis Based Book Retrieval Approach

The manner of keyword-based retrieval in today's libraries cannot achieve precise retrieval of books, which normally requires an understanding of the book contents. This paper proposes SemBRS, a retrieval system that enables book searching using queries regarding the key contents of books. In SemBRS, a multi-layer sequence-to-sequence model of semantic parsing is designed for analyzing queries in natural language to be structured retrieval statements. Relevant books are obtained based on the correlation of the parsed query information and the tags of books. As the fundamental work, the tags of each book are generated based on an analysis of the contents of the book using a multi-label classification model.