Multi-Criteria Decision Making And Machine Learning For Inventory Classification: A Case Of L'oreal

Abstract

For a long time, beauty has always attracted women, so there are many care and beauty products for women. However, in recent years, men's care and beauty have also become popular. Therefore, the beauty industry has always pushed forward with technology and launched many products for both sexes. Such mass launches have left many companies with some difficulty in calculating how much product they need to keep in stock. And L’Oréal is no exception to that difficulty. Therefore, this study comes out with the aim of making inventory calculation simpler and more accurate. This article will develop a method to effectively combine machine learning algorithms with multi-criteria decision-making techniques. MCDM has been around for a long time, but most of it is only used in selection, such as supplier selection. However, in the process of development, MCDM has recently been able to compute for ranking purposes. ABC Class is a popular method in inventory management, but most people use this method for two criteria: quantity and price. In fact, to calculate the amount to be stocked, there are many more factors. Therefore, combining MCDM and ABC classes will help the company evaluate more criteria. In this study, two main methods will be used: the Best-Worst Method (BWM) and VIKOR. The two methods are quite new and provide high accuracy. As for machine learning, we will use four quite popular algorithms: Support Vector Machine (SVM), Random Forest, Gradient Boosting and Artificial Neural Network (ANN).