Research On Opinion Leader Identify And Recommend Benefit Maximization Based On Community Detection

With the explosive development of Internet and e-business prosperity,personalized recommendation technology as users browse through efficiency and e-commerce means visitors to purchase rate more and more attention.The rapid development of technology in recent years,but still have low accuracy and the recommendation problem.To address these issues,many scholars,social networking and e-commerce integration,the "social recommendation",the concept of a branch of this field of research is the study,according to the two-step flow theory,opinion leaders,Word was introduced to strengthen the social recommendation.Firstly,social network analysis and recommended and summarizes the progress of the algorithm,and detail the two-step flow theory,social network analysis,community,discovery,personalization field of classical theory.Articles-through analysis research complement each other in various fields,has a fusion may,but realize the need to consider the impact on such problems.To this end,based on PageRank algorithm presented opinion leaders influence interference dynamic assessment recommending effectiveness and community recognition algorithm combining discovery and opinion leaders,design principle of effectiveness of community opinion leaders recommend sorting,opinion leader identification model is constructed.Finally,the simulation hypothesis and empirical analysis of validation,as well as opinion leader identification model validation.Results show that opinion leader identification model designed in this project than the traditional model has a wide range of recommended coverage and validity.