Machine Learning Algorithm for Intelligent Prediction for Military Logistics and Planning

Abstract

This paper compares various machine learning algorithms for predicting availability and possible reorder level of military logistics. As a case scenario, dataset of price of petroleum product was used to test the accuracy of the proposed algorithm. In most Military, facilities, machines and equipment relied heavily on the availability of petroleum products. Military logistics must be intelligent, based on informed deductions. Machine learning is now pervasive and is readily applied to various areas of life including the military. Result of the evaluation shows that artificial neural network (ANN)- 85.57% and logistic regression- 78.44% performed better than k-nearest neighbour (KNN)-74.98%, random forest(RF)-72.81% and Naive Bayes(NB)-74.85%. If used in Military logistics, there could be attendant benefit such as: accurate price policy formulation; proper budgeting estimation; meeting production and demand targets; proactive supply chain and value chain derivations; informed and intelligent decision making process; competitive advantage and continuous availability of supply critical to military; trigger of further research on innovative emergent technologies in this area, amongst other intangible benefits.