Data mining of Absentee data to increase productivity

Productivity of an organization reduces when employees are absent for prolonged duration. This could be avoided if the employee's absentia is understood in the early period and a substitute is assigned. This paper is a research to create classification model to predict whether an employee would be absent for short or long duration. Various classification models are studied and Multilayer perceptron in found to be the best suited model for this purpose. Absenteeism record of Courier company from UCI data repository is used for this study.