Artificial Neural Network Modeling of Job Satisfaction: A Case Study of ICT, Federal University of Agriculture, Makurdi

Abstract

Artificial neural network assessment and modeling of job satisfaction of the Information and Communications Technology (ICT) Directorate workers of Federal University of Agriculture, Makurdi was investigated in this study. Modified Nordic Musculoskeletal Disorder (NMDQ) questionnaires which incorporated health, safety and environment factors were used. The questionnaire consisted of a series of objective questions with 'yes', 'no' and 'I don't know' responses and some were multiple choice questions. Parameters such as health, safety, environment and ergonomic factors were obtained from questionnaires for the modelling of workers efficiency and job satisfaction. The efficiency of workers was determined and normal probability curve for the 40 workers was plotted to identify the outliers. The artificial neural network (ANN) modeling method was employed to predict job satisfaction using health, safety, environment and ergonomic factors as input parameters while job satisfaction was the output. Series of network architectures were considered using different training algorithms. The scale conjugate gradient SCG 4 [3-3] 2 1 was adopted as the suitable network architecture for predicting job satisfaction. Result indicated that the predicted values of job satisfaction were in the range of 1.42 -2.00 as compared with the actual values of 1.50 -2.00 obtained from the questionnaires. Statistical indicators of normal error (E), used for validation of the model gave minimal errors and varied in the range of -0.48 -0.08. The plot of the normal probability curve also indicated the presence of outliers or inefficient workers. Whereas most of the workers were satisfied with the existing health, safety, environment (HSE) and ergonomics (E) programs at the work place, some (outliers) were not. The presence of outliers calls for improvement of ergonomic conditions at the ICT directorate.

Latest Advancements in Solar Photovoltaic-Thermoelectric Conversion Technologies: Thermal Energy Storage Using Phase Change Materials, Machine Learning, and 4E Analyses

“…Some popular machine learning algorithms used for optimization include gradient descent, genetic algorithms, and reinforcement learning. By leveraging the power of machine learning for optimization, several novel researches have been conducted and findings published to aid the scientific community [7,[100][101][102][103][104][105]. There are three primary stages involved in an ML system (Figure 9).…”