| (英) |
We propose an approach using federated learning for predicting Wi-Fi and LTE transmission control protocol (TCP) throughput to reduce the delay between the output of prediction results and the problem of security risks by sharing the datasets, which is a problem with conventional machine learning methods. In this study, we constructed a machine learning model, implemented the federated learning model, and conducted prediction evaluation experiments using measured data using the federated learning method and the conventional method. From the experimental results, we show that the proposed method solves the conventional problems and achieves the same level of prediction accuracy as the conventional method. And, the prediction object range of the machine learning model was examined from the viewpoint of the problem assumed by the proposed technique. In addition, we proposed a method of transferring the learning model as a way to improve the prediction accuracy of the throughput in the proposed method, and evaluate the effect of this method. |