A Study of Mental Health Self-Monitoring Based on the Combination of BERT and Low-Code Platform

Abstract

This paper proposes a technical route for mental health self-monitoring based on BERT. This route can facilitate models that achieve good results on many tasks of natural language processing and can excellent in analyzing the emotions of the recorders. At the same time, the low-code platform, as an auxiliary system tool for software engineering, is able to deploy some machine learning tasks, including data preparation, storage, model building, etc., therefore, it used in this experiment to assist language models for sentiment analysis. With the combination of the two techniques, the accuracy of this technological route to facilitate sentiment analysis can reach up to 88.32%. And by reminding the changes in the emotions of the recorder, it can initially achieve the purpose of achieving mental health self-monitoring.