Estimation of Degree of Human Fatigue Score by Defined Data Distance in Large-Scale Data-Based Online Modeling

Abstract

It is important to quantify and prevent mental fatigue in order to forestall absence from work and death from overwork. In this study, we used "Large-scale Data-based Online Modeling (LOM)", one of the local modeling methods based on databases, to reproduce the characteristics of fatigue based on information derived from observation of biomedical signs. The subjects in our study were 10 male university students. After assuming a relaxed seated position for 10 minutes, they performed two repeated sessions of a mental arithmetic task for 30 minutes each session with a 30-minute break. Subjective symptoms of feelings of fatigue score were evaluated on a 10-point scale. We measured several biomedical signs and subjective symptoms of feelings of fatigue score. We propose that it is possible to estimate the degree of acute mental fatigue from biomedical signs using the LOM method. We propose a method for determining the reliability of estimating "degree of fatigue" by distance of neighboring data.