2D data arrangement to train ANN for depression levels measurement

TL;DR

This paper introduces a method using 2D data combining physical and psychological parameters to train ANNs for measuring depression levels more accurately.

Contribution

The novel contribution is the integration of psychological and physical parameters as a two-dimensional dataset for training ANNs in depression detection.

Findings

A significant correlation was found between increased stress perception and elevated heart rate and reduced sleep quality.

The dataset includes all levels of depression, enhancing the effectiveness of depression measurement using 2D data.

Using 2D data improves the ANNs' ability to detect interaction patterns between physical and psychological parameters.

Abstract

We arranged data to train Artificial Neural Networks (ANNs) designed as a depression-level measurement tool. Even though, as an advanced form of stress, depression impacts many physical parameters disorder, measuring depression using only physical parameters is insufficient. It is urgent to integrate comprehensively psychological and physical parameters as two dimensions, 2D, data. We harvested the dataset of 95 respondents from college students. The physical dimension consisted of four parameters measured noninvasively, and the psychological dimension was assessed using the Perceived Stress Scale (PSS). The initial analysis revealed notable correlations between increased stress perception and certain physical parameters analysis, particularly an elevated heart rate and reduced sleep quality. The highly significant p-value provided strong evidence that the observed difference in means is not coincidental. According to data processing, we have the data set including all levels of depression to enhance the effectiveness of measuring depression. Using two-dimensional data, we aim for the ANNs to learn interaction patterns between these parameters, improving accuracy in depression detection.

Figures