Abstract—The authors have been grouping and developing various applied sensing systems to solve difficulties in achieving agricultural technical teaching. For these purposes, the authors developed wearable systems containing the latest sensing modules, and measured acceleration data. Their basic data and data concerning subjects’ fatigue and subjects’ intensity of tasks were collected. The authors analyzed those data using statistics. Because thought that past methods and systems were insufficient to analyze human motion related to users’ fundamental information, their feeling of fatigue and of intensity about agricultural works. The authors have obtained and observed time-line hoe acceleration data when subjects have been cultivating using a common hoe. Subsequently, the authors have computed those data using basic statistical methods, as well as SEM (Structural Equation Models) and GM (Graphical Modeling) analysis based on multiple regressions. In future, the members hope to make useful suggestions to the field of agricultural informatics using those data. 
 

Index Terms—multiple regression analysis, SEM and GM analysis, correlation ratio, wearable sensing system, worker impression

Cite:  Shinji Kawakura and Ryosuke Shibasaki, "SEM, GM and Other Statistic Analyses Concerning Index Values Extracted from Outdoor Agricultural Workers Data," Journal of Advanced Agricultural Technologies, Vol. 4, No. 4, pp. 296-307, September 2017. Doi: 10.18178/joaat.4.4.296-307