A deep understanding of the factors affecting the characteristics of user experience helps to promote the benign development of China's mobile social network services, while how to improve the satisfaction of China's user experience has become a major key to the development of China's network commercialization applications. This paper draws on the application of machine learning algorithms in other areas of classification and identification, replicated to the classification and identification of user experience satisfaction to promote the high-quality sustainable development of mobile networks, and then through the analysis of the various factors affecting user satisfaction, to provide a basis for decision-making, so as to achieve an earlier and more comprehensive enhancement of user satisfaction. This paper adopts SMOTE oversampling for the unbalanced data of mobile user scoring dataset, and then adopts factor analysis for the main factors affecting customer satisfaction, and carries out Bartlett's spherical test. After that, GBDT model is established for user satisfaction prediction, and robustness test is used to optimize the model parameters, and the results show that the AUC value of GBDT model prediction is about 0.99. The time complexity of the algorithm can be greatly reduced and is more acceptable to industry.
Read full abstract