Critical Study of Machine Learning Models for Retail Sales Prediction

-, Athul Ajayghosh (2025) Critical Study of Machine Learning Models for Retail Sales Prediction. Masters thesis, Dublin, National College of Ireland.

Preview	PDF (Master of Science) Download (987kB) | Preview
Preview	PDF (Configuration Manual) Download (349kB) | Preview

Abstract

Proper sales projections of retailing is essential in minimizing inventory expenses and enhancing efficiency. The traditional models like the ARIMA and SARIMA have been found to be useful in predicting the short run but they do not cope with the non-linearity and dynamism of the current retail data. The latest developments in the field of forecasting have presented machine learning algorithms such as Random Forest, XGBoost and CatBoost, as well as deep learning based algorithms such as Long Short-Term Memory (LSTM), Gated Recurrent Units (GRU) and hybrid Models. More recent scalable models shows good long-horizon accuracy, but there are few studies which report their performance in terms of accuracy, efficiency, and reduction of inventory costs in retail sales forecasting.

The paper compares fourteen forecasting models of the weekly revenue of the UCI Online Retail data based on a constant four-week prediction horizon. The classical models had moderate accuracy (ARIMA RMSE 63796; SARIMA 57585), whereas the machine learning models enhanced their accuracy by non-linear learning of features. The best results were obtained with the method of deep learning, and the lowest error was obtained with CNN-LSTM (RMSE 31948; MAPE 8.2%).These results underscore the usefulness of sophisticated neural models in enhancing sales prediction and minimizing inventory-related expenses and offer retailers some practical ideas and empirical data to the current forecasting literature.

Item Type:	Thesis (Masters)
Supervisors:	Name Email Sahni, Vikas UNSPECIFIED
Subjects:	Q Science > Q Science (General) > Self-organizing systems. Conscious automata > Machine learning H Social Sciences > HD Industries. Land use. Labor > Specific Industries > Retail Industry
Divisions:	School of Computing > Master of Science in Data Analytics
Depositing User:	Ciara O'Brien
Date Deposited:	04 Sep 2026 11:44
Last Modified:	04 Sep 2026 11:44
URI:	https://norma.ncirl.ie/id/eprint/9840

Actions (login required)

View Item