CNH Markov Chain-Based Forecasting Model

Abstract

The purpose of this thesis is to develop a Markov Chain-based forecasting model that can accurately predict the future demand quantities of service parts for CNH Industrial. Markov Chain transitional probability tables will be created for each part. These will be used to create systems of linear equations that could be solved to derive the long run probabilities of each demand state possibility. These long-run probabilities will be used to determine the expected value of the future demand for each part. Finally, these expected values will be compared to the actual demand of the period being predicted and the accuracy for the sample will be compared to the accuracy of currently existing algorithms to determine if there would be value in attempting to incorporate this model into the existing forecasting system.