An Integration of Requirement Forecasting and Customer Segmentation Models towards Prescriptive Analytics For Electrical Devices Production

Abstract

Material requirement planning is an essential role of a manufacturing business. Manufacturers need to find an effective way to manage material planning among the changes. This research is designed to create an integrated model of time series purchasing forecasting model and customer segmentation model in electrical equipment procurement for risk assessment and prescriptive model building. The methods used for forecasting are compared between Gradient Boosted Tree (GBT), Artificial Neural Network (ANN) and Decision Trees (DT) while the K-Means Clustering is selected to segment customers optimally. Henceforth, customers can be classified into three groups; Good, Moderate and Normal. The results of both methods are then used to generate a risk assessment matrix. Finally, the researcher analyze with the prescriptive analytics driven by the evolutionary optimization method to create a strategy and allocate parts which align to customer behaviour and according to the company policy.