Hybrid Recommendation Algorithm based on Product Popularity and User Preference

Abstract

In today's society, recommendation system is widely used to recommend users' preferred products and content. However, with the rapid development of the Internet, there are more and more contents on the Internet, and the algorithm complexity of recommending products to users through the recommendation system is getting higher and higher, and the recommendation algorithm is also more and more inclined to recommend popular products to users, which also leads to the decline in the novelty and diversity of the recommendation algorithm. In order to solve this problem, the study first adopts the method of inhibiting the recommendation ability of current popular products, and on this basis, introduces the influence of user rating on the recommendation ability of different products. At the same time, the product is classified, take the movie for example, can be classified as action film, comedy, children's film and so on. There is a very high probability that an adult will not like children's films, and a very high probability that a child will not like horror films. Impose a penalty on a product category that the user may not like, reducing the likelihood that the product will be recommended to the user. The proposed algorithm is compared with other excellent algorithms in this field on MovieLens movie score dataset, and the results of the experiments showed that the algorithm improved in accuracy, diversity, and novelty.