Title: Construction of mental health monitoring system based on model transfer learning algorithm

Authors: Panpan Li; Feng Liang

Addresses: College of Healthy Management, Shangluo University, Shangzhou District, Shangluo, Shaanxi Province, China ' College of Healthy Management, Shangluo University, Shangzhou District, Shangluo, Shaanxi Province, China

Abstract: In order to monitor people's mental health in real-time and effectively, this study has conducted in-depth research on the model transfer learning algorithm, including its learning process, classification criteria, network structure optimisation, etc. The research takes model transfer learning algorithm as the main research method, and innovatively adopts residual learning and gradient descent algorithm to optimise the performance of model transfer learning algorithm, and then compares and analyses the application effects of model transfer learning algorithm and traditional machine learning algorithm in various data sets of mental health monitoring, so as to ensure the accuracy of monitoring results. The results show that the model transfer learning algorithm is significantly better than the traditional machine learning algorithm in accuracy, recall and F1 score, and it requires less network training time. This shows that the mental health monitoring system based on model transfer learning algorithm has good performance and can monitor mental health accurately and efficiently.

Keywords: model; transfer learning; mental health; monitoring system.

DOI: 10.1504/IJWMC.2023.129088

International Journal of Wireless and Mobile Computing, 2023 Vol.24 No.1, pp.58 - 65

Received: 02 Apr 2022
Received in revised form: 05 Jul 2022
Accepted: 26 Aug 2022
Published online: 17 Feb 2023 *