Descriptive analysis of online roulette gamblers: segmentation of different gamblers based on their behavior using data mining algorithms

Abstract

The popularity of gambling activities has been increasing over the last decades, with onlinebased gambling being a key driver of its growth due to the ease of accessing online platforms. Consequently, there is a severe concern that the negative social impact of gambling arises, and regulatory agencies are identifying and managing those effects. In this context, a potential solution to address those effects is based on the concept of 'Responsible Gambling', which means playing consciously, with complete control of time and money. The present study aims to segment online gamblers based on their playing behaviors, differentiating groups as much as possible and ultimately identifying a cluster with players of concern. This is achieved using unsupervised learning algorithms such as K-Means, Hierarchical Clustering, or Self-Organizing Maps. The information on which this project is based reflects the activity on some of the Portuguese online gambling platforms over 2019. Available data covers multiple aspects such as the gambling institution, type of gambling, player identification, each player's total bets, and the following outcomes of it.