Satisfaction Prediction for Heterogeneous Mobile Network Cells Based on Semi-supervised Transfer Learning

Semi-supervised and transfer learning are efficient methods to improve performance when labeled data is challenging to obtain. With the development of mobile communication network, the cellular mobile network system is becoming more and more complex, and there are more and more factors affecting the user satisfaction of mobile network. In order to solve the satisfaction optimization problem in the heterogeneous mobile cellular network, this article proposes a satisfaction prediction model for heterogeneous mobile network cells based on semi-supervised transfer learning (SP-SSTL). Firstly, SP-SSTL uses semi-supervised to train a model based on the source domain, which is user data. Then uses transfer learning to train a model based on the target domain, which is cellular network cell data. Meanwhile, the transfer learning model is fine-tuned using domain adaptation. Experimental results on different domains show that semi-supervised learning can significantly improve training efficiency and accuracy when labeled data is challenging to obtain; moreover, transfer learning can solve the problem of insufficient data volume in the target domain. Compared with the existing methods, our proposed method improved the accuracy by 15%, and the efficiency also significantly improved.