Machine Learning in Demand Forecasting - A Review

Abstract

Demand forecasting is of great importance in many industries such as agriculture, electric power, tourism, retail sales and manufacturing companies, etc. It plays a vital role in the decision making of every business. This paper survey various state of art methods on demand forecasting with a focus on machine learning. The machine learning techniques have been categorized into three categories namely time series analysis, regression-based methods and supervised/unsupervised models. The pros and cons of the various machine learning techniques are discussed and their performance measures are compared. The comparison conceals that LSTM has a notable result, but its computation time is higher than any other method. Another field of future research includes regression-based methods, hybrid models and ensemble models. This study gives the reader an idea of demand forecasting in the field of machine learning.

Keywords: Demand Forecasting; Machine Learning; LSTM; Regression-Based Methods; Hybrid Model; Ensemble Model