Recommendation Algorithm Combining Weighted Average and Trust Degree

Most of the traditional rating forecasts only rely on the user’s rating of the project, and there is no integration of user’s social relationship. Therefore, this paper proposes a recommendation algorithm combining weighted draw and trust, which combines explicit features such as users’ ratings of projects with implicit features such as the trust relationship under the premise of considering users’ social relations. In this way, users’ preferences can be obtained more accurately, and the problem of cold start of users caused by sparse data can be alleviated. In addition, the trust degree proposed in the article is the synthesis of multiple relationships, which is manifested in the subjective trust of users to trusted users and the influence of trusted users among all users. Through experimental analysis, it is superior to other baseline models in three commonly used data.