Research And Realization Of Recommendation System Based On Trustworthiness

With the rapid development of Internet and E-commerce, there are more and more data online. How to filter useful data from the huge database is becoming more and more important. Recommendation system provides an effective means to solve this problem. The most popular personalized recommendation system can give different people personalized recommendation results. In the personalized recommendation field, the most commonly used algorithm is collaborative filtering algorithm. Collaborative filtering algorithm can give target user recommendation results according to the evaluation of his similar users. However, the way to find similar users of the traditional collaborative filtering algorithm is single and has the certain limitation.This paper presents a new recommendation system. In this system, trustworthiness and collaborative filtering algorithm are combined. Trustworthiness is used to replace similarity in the traditional algorithm as the standard of finding target user’s neighbors. This paper gets more comprehensive and accurate trustworthiness data from subjective and objective aspects. When talking about interest factor, time factor is considered for aided analysis. This recommendation system based on trustworthiness can get better recommendation results. In addition, for multiple users’ recommendation, this paper proposes a new model, which is comparing with the based user to calculate the result. The new model can effectively reduce calculation time and improve system performance.In order to verify the feasibility of the system, this paper uses python language to realize the system. In order to verify the effectiveness of this paper, three different groups of contrast experiments are designed on the selected data set. The experiments show that the recommendation system based on trustworthiness can effectively enhance the quality of recommendation system and get better recommendation results.