A Comprehensive Study On Key Factors Of Encouraging User Participation In Social Media

As technology has evolved and the accessibility of online media has increased,the popularity of social media has grown at an unprecedented rate. Social mediaprovides a new channel of communication, and gradually changes the way peopleconsume information. Among all, the most valuable components of social mediaare participation and activity. The unique characteristics such as social interactionand information flow decide the great values of social media. Thus, an in-depthunderstanding of user participation can help implement social media into businessmarketing plans, and beneft the life for civil society worldwide.In this paper, we conduct a comprehensive study on user participation in socialmedia. In order to increase user experience and motivate user engagements, we con-centrate on three key dimensions: diversity to help users explore various territory,serendipity to bring users more surprise and interestingness, and social influencefrom friends that increases users’ likelihood to respond. Our study towards eachdimension are elaborated as below.First of all, we have all learned one of life’s great lessons: diversity is beau-tiful. In the study of diversity, we propose four metrics to quantify diversity inmicroblogs, and the results of temporal analysis show that diversity increases with time. Our experiments also demonstrate that diversity is a strong predictor of userparticipation including social interaction and user engagement.In the study of serendipity, we investigate the ubiquitous occurrence of serendip-itious information difusion and its efect in the context of microblogging commu-nities. Serendipity is a benefcial discovery that happens in an unexpected way. Ithas been found spectacularly valuable in various contexts, including scientifc dis-coveries, acquisition of business, and recommender systems. We refer to serendipityas unexpected relevance, then propose a principled statistical method to test theunexpectedness and the relevance of information, which identifes a serendipitousdifusion of information. Our fndings based on large-scale behavioral analysis revealthat there is a surprisingly strong presence of serendipitous information difusionin retweeting, which accounts for more than25%of retweets in both Twitter andWeibo. Upon the identifcation of serendipity, we are able to conduct observationalanalysis that reveals the beneft of serendipity to microblogging users. Results showthat both the discovery and the provision of serendipity increase the level of useractivities and social interactions, while the provision of serendipitous informationalso increases the influence of Twitter users.In the study of influence, for the purpose of encouraging overall participationthrough social influence, we formulate the problem of participation maximizationand propose an advanced algorithm as solution. More specifcally, we study thefollowing three important aspects of the problem:1) influence validation, for whichwe propose a statistical shufe test to verify that social influence does motivateposting behaviors in some forums;2) influence analysis, for which we propose an it-erative algorithm to estimate influence probabilities between every pair of users; and3) participation maximization, for which we propose a budget constraint SideBarscheme to allocate threads to all users based on their social influence to maximizeuser participation, and we formulate the problem into a social welfare maximizationproblem with submodular utility functions and apply an approximation algorithmTABI to solve the problem.To sum up, in this paper, we conduct in-depth analysis to explore three key factors of user participation in social media: diversity, serendipity and influence.Our study show that all the three dimensions play signifcant roles in encouraginguser participation. Thus, in applications and services of social media, such as recom-mendation, retrieval and advertising, practitioners could make use of the proposedmethods and algorithms to incorporate the three factors and in turn beneft fromincreased participation.