Data Mining Techniques Applied to a Manufacturing SME

This paper examines how data mining, an aspect of analytical science, can be applied to assist a Small to Medium Enterprise (SME) industry using unsupervised learning techniques, association rules and time-series analysis. Whilst recent developments have meant it is now possible for SME to compile large amounts of commercial data, this information is rarely utilised effectively. The study builds on a number of standard data mining techniques to produce a tailored set of analyses that provide maximum benefit to the company. Self-Organising Maps were utilised to visualise the core characteristics of the firm's customers. The study outlines a new technique to determine associations between customer variables using the arules package available within RStudios. Finally, time-series forecasting was conducted highlighting the seasonal variations and trends for potential growth in the coming year.