Book recommendation method, device and equipment based on collaborative filtering and TransH

The invention provides a book recommendation method, device and equipment based on collaborative filtering and TransH. The method comprises the following steps: acquiring book score data corresponding to a target user; determining similar users corresponding to the target user based on the book score data, and determining a first book recommendation result according to the book score data corresponding to the similar users through a collaborative filtering algorithm; determining a second book recommendation result based on a pre-constructed book semantic similarity matrix, the book score data and/or the first book recommendation result; wherein the book semantic similarity matrix is constructed based on a target TransH model and a book knowledge graph; and fusing the first book recommendation result and the second book recommendation result to obtain a mixed book recommendation result corresponding to the target user. According to the method, the book recommendation list can be more reasonably and accurately