Research of Personalized Recommendation Algorithm based on Similarity and Diversity

Now the general recommendation system only recommends high similarity products, and if only the similarity is likely to cause the waste of resources and the unsuccessful recommendation, for example, a user likes to use WeChat. [1] Based on the similarity, it is possible to recommend the flying letter to him, so the function of the product is too similar and will cause it. Due to the waste of resources, the difference should be considered properly in the design of recommender systems. The idea of this paper is to complete the research of product attribute similarity algorithm first, and to complete the design of recommendation mechanism based on similarity and difference. Finally, the accuracy and recall rate of various recommendation mechanisms are compared by experiments to determine their superiority.