Collaborative Filtering Recommendation Algorithm Fuses Semantic Nearest Neighbors Based on Knowledge Graph

To solve the problem that collaborative filtering only uses the items-users rating matrix, a collaborative filtering recommendation algorithm fuses semantic nearest neighbors based on Knowledge Graph is presented. This method fuses semantic nearest neighbors and embeds the existing semantic database into a low-dimensional semantic space. It combines the semantic information of items and collaborative filtering using calculating the semantic similarity between items. The shortcoming of collaborative filtering which does not consider the semantic information of items is overcame, and therefore the effect of collaborative filtering result is improved on the semantic level. Experimental results show that the proposed methods can get higher values on Recall, Accuracy and F-value for collaborative filtering systems.