Research On Intelligent Recommendation System Based On Trust Deliver

In recent years,with the rapid development of information technology and Internet technology,the amount of information has shown an exponential growth.So how to find users' needs from the mass of information has gradually attracted people's attention.For example,in the Internet era,e-commerce websites are developing rapidly,but users are faced with the problem of how to choose commodities among lots of commodities.The emergence of recommendation system can solve this problem.In the most common recommendation system,the main principle is to collect the items or information which the user is interested in,then classify the items or information,and finally recommend the same kind to the user.Recommendation system can not only save users' selection time and improve users experience,but also bring more benefits to e-commerce websites.Firstly,this thesis introduces the current recommendation systems,including collaborative filtering-based recommendation algorithm,content-based recommendation algorithm,knowledge-based recommendation algorithm,trust-based recommendation algorithm and hybrid filtering-based recommendation algorithm.Then,it introduces the problems and challenges of common recommendation systems in detail,such as cold start problem,data sparsity problem and so on.In addition,it systematically introduces the common evaluation indicators in recommendation system,such as absolute average error,coverage and so on.Then,from the user's perspective,it proposes a trust-based prediction approach for recommendation system,which uses social networks with trust ratings to generate recommendations for users who trust them.The trust value between users is defined,and the trust value is calculated by quantifying the trust relationship between users to predict and score the unrated items,so as to alleviate the problem of cold start and data sparsity.In addition,from the item's perspective,it combines the decision tree algorithm in machine learning with the content-based recommendation system to form a content-based intelligent recommendation method.Using the decision tree algorithm,the classification of items is more accurate,so as to further alleviate the cold start problem of items and bring better recommendation effect.Finally,the trust-based prediction approach for recommendation system and the content-based intelligent recommendation method are combined to form an intelligent recommendation method based on trust delivery.Experiments are conducted to verify theeffectiveness of the proposed method.