UGC Quality Predictive Method Based On The User Reputation Rating

Web2.0 era has changed the original mode of information production,before the Web2.0 era,the information is generated by professional editors after proofreading the books,newspapers,TV programs,however,the current information can be generated and wantonly spread by any individual or groups.The surge of UGC(User-Generated Content)enriches the network information resources while the quality of UGC is uneven which affects the development of network.Evaluating and discriminating the low quality information timely and effectively,organizing and managing the high quality information effectively,both the two aspects affect the survival and development of the network information environment.In recent years,there are researches about quality evaluation and control of the UGC,most of which are from the perspective of information content,ignoring the factors of users' behavior.This research develops from the perspective of user information behavior,by mining the relationship between user's historical behavior and UGC quality,building user reputation scoring mechanism,thus establishing the UGC quality pre-judgment model based on the user reputation rating and realizing the effective forecasting the users' output quality of UGC in the future.There are six chapters in this research.The first chapter and the second chapter mainly explain the background and some related theories of the UGC research.The third chapter and the fourth chapter mainly analyze the user information behavior factors which affect UGC quality and build the prejudgment model of UGC quality based on the user reputation rating by mining and analyzing the creation,transmission,comments and other historical behaviors of the UGC from the past information activities so as to set up the personal information behavior of credit rating model for the user.The fifth chapter and the sixth chapter mainly verify the effectiveness of the model and look forward to the future research.In order to verify the validity of the model,this study obtains six users' micro-blog behavioral data by the method of crawler and the users' submission and makes a week as total evaluation cycle.The users' information behavior data are obtained under the 53 points in time and calculate the users' reputation scores.At the same time,in order to reduce the subjectivity of the evaluation process,the UGC information quality evaluation system is designed to assist in evaluating the users' information quality from historical behavior data.The results show that the model is different from the previous method to evaluate the quality of UGC which depends on single text analysis.It combines the methods of information science and behavioral science,which has high accuracy for prediction.However,due to the lacking of unified user reputation rating scoring rules,there is a certain degree of subjectivity in this research and it can only make a pre-judgment of the possibility of the users' future behavior: good or bad,based on the user's reputation from their past behavior records.Therefore,the future research will further explore users' behaviors that affect UGC quality based on unified scoring standards,combining it with semantic content analysis and emotional analysis to improve the accuracy of UGC quality prediction.This research is supported by the National Social Science Foundation of China,and the project number is 15BTQ064.