Exploiting And Motivating User Generated Content: Competitive Intelligence Modeling And Economic Incentive Exploration

Information technologies have been continuously evolving to change our lifestyle,with more and more user generated content emerging on the internet,such as query logs,word-of-mouth,question and answering.This dissertation conducted research from two persectives regarding user generated content—the exploration and application of its value,and behavioral analyses on its generation: 1)Representing the collective intelligence,user generated content contains valuable businesss information.Via mapping competitors’ associations in user generated content,the first part of this dissertation focuses on competitive intelligence modeling,to enhance user generated content based business analyses.2)Because of user generated content’s value,some platforms attempt to stimulate its contributions through economic incentives.Considering the pro-social motivation in users’ generation process and the motivation crowding-out effects,the second part of this work aims to reveal users’ contribution patters under incentivization.In competitive intelligence modeling,a bipartite graph based approach is developed to analyze businesses’ competitive degrees and a topic based keyword recommendation method is designed to help achieve competitive advertising.Usually,market competitors share common attributes,which project to the associations between attribute keywords and different business entities in user generated content.The bipartite graph approach exploits exactly such hidden associations to measure the competitive degrees between businesses.A series of experiments demonstrate that this approach could help effectively improve competitior ranking and market share prediction in practice.By combining the association information with topic modeling in user generated content,a keyword recommendation method is further designed to help better achieve competitive advertising.Observing the attempts of using economic incentives on quite a few platforms,this dissertation specifically discusses the pro-social motivations in user generated content and how economic incentives could crowd out such internal motivations,leading to continuous spillover effects.In the context of review writing,a quasi-experiment is created through Propensity Score Matching and Difference-in-Differences,where we specifically analyzed how reviewers would change their behavior after previously having experiences of accepting economic incentives.We also designed a series of behavioral experiments to explore strategies that can help undermine the crowding-out effects of monetary incentives on pro-social behavior.Experimental results confirm that goalsetting and challenge-seeking are effective to achieve this purpose when they are used together with economic incentives to stimulate pro-social behaviors.These research findings provide theoretical and practical guidelines for policy design of platforms which rely on pro-social contributions.