Mining Automatic Teller Machines (ATM) - Transactional Data

Abstract

Understanding hidden patterns underlying transactional databases can help solve some of the bank's business problems and also present new business opportunities. Data mining and knowledge discovery in databases have been attracting a significant amount of research, industry, and media attention of late. Time series data is perhaps the most frequently encountered type of data examined by the data mining community. Clustering is one of the frequently used data mining algorithm, being useful in it's own right as an unsupervised exploratory technique, and also as a subroutine in more complex data mining algorithms such as rule discovery, indexing, summarization, anomaly detection, and classification. Various methods of analyzing data for patterns were reviewed. This study showed that there are interesting patterns in Automatic Teller Machine (ATM) transactional data in form of clusters and trends when the data was translated into a time series. The characteristics of clusters were described. The study however was limited by the breadth of data available. It recommended that transactional data cannot be investigated alone. There is need for transactional data to be argumented with other data such as personal information and other relevant data in a warehouse so as to more meaningfully describe the patterns existing in transactional