도서 정보 및 도서 리뷰 데이터를 활용한독서 수준 기반의 추천 모형 개발

Abstract

This study presented a book recommendation model based on reading level which can be applied to the publishing industry. Based on the converged technologies of IT industry and AI industry, book level was calculated based on text difficulty analysis, and the reviews were analyzed using NLP techniques and artificial neural networks based on clusters based on customers’ characteristics. The derived book level was set as a pre-score and the expected level of book experience of target customers derived through a fully connected neural network was set as a post-score. From those scores, we derived a model of book recommendation as a rank considering customers’ reading level. From a survey conducted to verify this mdoel, the model had a statistical significance of correlation with survey results. In addition, this study presented a reading roadmap and business model considering reading level as a way to apply this model developed in this study.