Designing a Framework for Predicting Employees Absence Through Artificial Neural Network Approach

Absence from work is an important challenge for industries and organizations. This costly problem of the organizations has caused managers to think of ways to decrease it. The first step in decreasing the cost of absence relates to identification of the factors affecting absence. The present research, which is the first one using approximate nonlinear functions of neural network approach, presents an appropriate model for predicting the amount of absence and factors affecting it. To do so, 14 effective factors were identified as input parameters, and then the amount of absence was predicted by multilayer perceptron model, through which the manager can decrease the amount of absence. Afterwards, using sensitivity analysis, the effect of each of the input parameters on the outputs of this model was analyzed. The final results indicated that the parameters of organizational commitment and job diversity were mostly effective on absence