CLASSIFICATION AND IDENTIFICATION OF THE CAUSES OF ABSENTEEISM IN A PUBLIC TRANSPORT COMPANY USING CLUSTER ANALYSIS AND PRINCIPAL COMPONENTS TECHNIQUES

Abstract

Absenteeism is the practice or custom of an employee to be absent from workplace. Its causes are diverse and may affect the workers income as well as to cause operational disruption, stress the administration and also financial losses for the company. Cluster analysis is a multivariate tool that can be used to determine groups in the sense that each group has its own characteristics in terms of the observed variables. In this sense, that technique can be used as a support to show which characteristics may contribute to absenteeism. We use the Ward hierarchical algorithm to build the clusters and to compare the groups the Kruskal-Wallis nonparametric test is adopted. Finally, a study on the strength of association among the variables is developed using Spearman’s correlation and for the relationship among those variables related to absence and social aspects, we use the principal component analysis. Moreover, the study indicates the possibility to determine three heterogeneous groups in the company and to show characteristics in those groups which are potential factors that cause absenteeism to a greater or lower extent.