Research On Quality Evaluation Of Mobile Internet Video UGC Based On Classification Algorithm

With the rapid development of Internet, the Web2.0 times is gradually mature in which the content is mostly generated by users. In recent years, the development of mobile Internet is to bring great changes to people's lives. UGC is a new content generation and organization form in the Web 2.0 environment, which has attracted much attention. At present, video sharing site, microblogging, blog, Q & A community are mainstream forms of mobile Internet UGC business. Mobile Internet has injected new impetus into the development of UGC. In recent years, the number of user-generated content has increased rapidly. However, more and more quality problems have been exposed, and its overall quality has yet to be improved. At present, although the number of user-generated content is increasing rapidly, the overall quality problem remains to be improved. The quality of UGC can be improved by choosing the scientific evaluation method to improve the environment of the UGC-based network platform. Based on the user-generated content quality evaluation, reasonable incentive measures can help users generate more quality content.In this paper, the mobile Internet video UGC is selected as the research object, and the quality of the video is evaluated by mining the video data based on classification algorithm. Based on the previous research and the characteristics of video UGC, the quality evaluation framework including object layer, dimension layer and measurement layer is constructed. The object layer includes video production level,video content itself, video viewing experience, video content utility to make a comprehensive and scientific evaluation. Based on the dimensionality index, the dimensionality index of the dimension layer is designed. According to the quantification index of the dimension layer,the video is scored manually. The principal component analysis method is used to determine the index weight. The quality of video UGC based on principal component analysis (PCA) is obtained by manual scoring and index weighting, and then the video quality is classified into high and low quality. The mode of content and user interaction is constructed, from which the index of measurement layer are taken.The results of quality classification based on principal component analysis and the data of measurement layer constitute the quality evaluation model based on classification algorithm. The sample data that we apply to this model are divided into training sample set and test sample set. The trained sample sets are used to train the model, and then the model is used to predict the quality of the test sample set. The results show that the model is highly operable and scientific.In this paper, we choose the video UGC from the UGC channel of Youku App to do the empirical analysis. The data of 892 videos are captured, and the quality classification of these videos is obtained by questionnaire and principal component analysis. The data of measurement layer index and video quality classification based on principal component analysis together constitute the sample data. The sample data are divided into the training sample set and the test sample set after the sample balance processing of the sample data. Taking the test sample set as an example, the quality evaluation model based on the C5.0 classification algorithm has 94.62% accuracy of classification and prediction of video quality. Finally, this paper compares the error and profit of the four classification algorithms. The results show that the C5.0 algorithm has the best prediction accuracy and the best forecasting profit.