Research On Personalized Recommendation Methods Based On Motivation,Trust And Chance

With the advent of the information age,social networks have gained popularity.The network includes a large number of valuable resources.However,as the number of users increased,the data released and shared also exploded,causing information overload.In order to solve the problem of information overload,the recommendation algorithm has been used to the Internet applications.However,in a large social networking platform,the traditional recommendation systems have many limitations,which lead to inaccuracy and homogeneity of the recommendation results.Based on the three elements of social network(motivation,trust and chance),this paper proposes a personalized recommendation algorithm based on motivation,trust and chance to construct a recommendation system suitable for social networks.First of all,this paper analyzes the motivation of users to post in social networks,and divides them into two categories: access information and social interaction.For the motivation of accessing information,we extract the user's needs from the posted content,and carry out motivation-based personalized recommendation by calculating the topic relevance.Then,for the motivation of the user's social interaction,we analyze the social relationship of the user from the relationship information of the user,calculates the trust degree between the users through the user-generated content information,the interaction information and the community group information,thereby discovering the relationship worthy of the user trust,and then performing the trust-based personality.Recommended.Finally,the traditional recommendation system only focuses on accuracy and ignores the limitations of diversity,as well as the long tail and recommended homogeneity of the recommendation system.This paper applies the chance theory to the recommendation system to construct a chance-based personalized recommendation algorithm.By linking the weak signal requirements with the implicit related resources,the potential needs and preferences of the users can be explored.This paper adopts the dataset of Sina Weibo.In the experimental design,the individual recommendation models are constructed for the three factors of motivation,trust and chance,the three models are independently tested to verify the effectiveness.Finally,the above three models are combined to construct a personalized recommendation method based on motivation,trust and chance.A large number of experiments and comparing verifies the feasibility.The experiment also adopted a variety of evaluation indicators,and achieved good recommendation quality in each evaluation index,which proved that the proposed algorithm has a good recommendation effect.