Collaborative filtering recommendation algorithm based on weighted item category
Jie Qin, Lei Cao, Hui Peng · 2016
User model is the key of recommendation algorithm, in order to solve the great sparsity of user-item rating model, proposed a collaborative filtering recommendation algorithm based on weighted item category. The algorithm evaluates the rated items by category, and with weighted scores summation, which changed the user-item high-dimensional rating data into user-category low-dimensional statistical data, and it was used as user feature model. Based on the model, forgotten function and user attribute information were added in, optimized the user model two times. Experimental results show that the proposed model is able to respond the sparsity of user-item rating problem, compared with cloud model, MAE was improved by 2%, the optimized model even has better quality, and which can solve cold-start problem.