Regression Analysis for ATM Cash Flow Prediction

One of the most challenging task for a bank is to maintain cash in their ATMs (Automated Teller Machines) so that they can easily serve their customers. To solve this problem, they create a daily estimate for each of their ATM, which can result into “Out of Cash” or “Over Stock” situations. This requires a solution which can resonably predict how much cash inflow would be needed for the next day by examining and learning from past transactional data. We present results for regression techniqes, including using the LSTM model for time-series for the first time to the best of our knowledge, to solve the “Cash Estimation” problem. This would allow banks to adopt to the changing needs of cash according to specific occasions, holidays, etc. The dataset that we would be using is transaction record for past 2.5 years for ATMs situated in a busy district of Karachi, Pakistan. This research will help banks in effectively reducing extra cost error which they bear for maintenance of their cash as well as increasing customer satisfaction.