International Journal of Computer Applications (0975 – 8887) Volume 133 – No.3, January 2016 13 An Artificial Intelligence ATM forecasting system for Hybrid Neural Networks Renu Bhandari M.Tech Student, RIMT-IET Mandi Gobindgarh, Punjab Jasmeen Gill Assistant Professor, RIMT-IET Mandi Gobindgarh, Punjab ABSTRACT Automatic teller machine (ATM) is one of the most popular banking facilities to do daily financial transactions. People use ATM services to pay bills, transfer funds and withdraw cash. Accurate ATM forecasting for the future is one of the most important attributes to forecast because business sector, daily needs of people are highly largely dependent on this. In recent years, Neural Networks have become increasingly popular in finance for tasks such as pattern recognition, classification and time series forecasting. Every financial institution (large or small) faces the same daily challenge. While it would be devastating to run out of cash, it is important to keep cash at the right levels to meet customer demand. In such case, it becomes very necessary to have a forecasting system in order to get a clear picture of demand well in advance. In this research article an integrated BP/GA technique is proposed for accurate ATM forecasting. The results are very encouraging. The comparison of proposed technique with the previous one clarifies that the proposed model outperforms the previous models. General Terms Hybrid Neural Networks, Artificial Intelligence. Keywords ATM Forecasting, ANN, Back propagation Algorithm, Genetic Algorithms, Hybrid Techniques. 1. INTRODUCTION ATM is a computerized telecommunication device that provides a financial institution's customers a method of financial transactions in a public space without the need for a human clerk [2]. Most ATMs are connected to international bank networks, enabling people to withdraw and deposit money from machines not belonging to the bank or country where they have their account. Automated Teller Machines are one of the most important cash distribution channels for the banks [11]. Bank customers are able to do their regular transactions such as bill payment, transfer funds and cash request without worrying about banks’ working days, business hours, etc. One of the primary concerns with ATM management is to determine the efficient level of cash inventory in each machine and there are different techniques offered to handle this problem [3]. Forecasting is a phenomenon of knowing what may happen to a system in the next coming time periods. For financial institutions, providing an adequate supply of currency to meet customer needs is a continual challenge. Running out of cash at an ATM or other location means reduced revenue from lost surcharge fees and increased expenses due to emergency currency deliveries. But overstocking currency means banks can’t invest this non-earning asset to generate interest income [10]. The ability to predict the future demand estimate of cash for automatic teller machines (ATMs) called ATM forecasting. The primary objective of ATM forecasting is to ensure that cash is used efficiently and effectively throughout the branch network [9].ATM forecasting and services availability is one of the most important factors in the ATM network services business. Using ATM cash management optimization and efficient cash loads routing, banks can avoid of stuck ATMs with cash and manage the system in dynamically changing environment by achievement the different requirements of ATM network participants. Recently, more banks are turning their attention to derive greater efficiency in how they manage their cash at ATMs [2].The key to the ATM’s forecasting algorithms is to capture and process the historical data such that it provides insight into the future. Newly, some authors attempted to optimize the cash by modeling and forecasting the demand [3]. However, the high variance and non-stationary of the underlying stochastic cash demand process can affect reliability of such approaches. Furthermore, the demand of cash is not only influenced by time, but it follows different tendencies that make modeling even more difficult. For example, holidays, weekends, starting of month, festival days etc. [2]. The remainder of the article is organized as follows. Section 1 introduces the ATM forecasting. Next, a brief description of ANN, BP and GA is given in Section 2, Section 3 and Section 4 respectively. The details of the integrated BP/GA technique for ATM forecasting model are shown in Section 5, followed by results and discussions in Section 6. Finally, Section 7 contains the summarized conclusions. 2. ARTTFICIAL NEURAL NETWORKS Artificial neural networks provide a methodology for solving many types of non-linear problems that are difficult to solve by traditional techniques. Essentially, an artificial neural network can be defined as a pool of simple processing units (neurons) which communicate among themselves by means of sending analog signals. These signals travel through weighted connections between neurons. Each of these neurons accumulates the inputs it receives, producing an output according to an internal activation function. This output can serve as an input for other neurons, or can be a part of the network output. Neural Networks have three building blocks- Learning Mechanism, Neural Network Architecture and Activation function [14]. 3. BACK PROPAGATION ALGORITHM One of the most popular training algorithms in the domain of neural networks used so far, for ATM forecasting is the back propagation algorithm (BPN). It is a gradient descent method. The algorithm suffers from several problems, like the local minima problem, scaling problem, non-suitability for complex problems [13].