The mathematical methodology most often used by automatic cash forecasting tools is based upon comparative statistics. Such statistics are compiled using various long-standing linear regression techniques that depend on repeating patterns within data.
Cash use and demand is the result of many undetermined factors, including seasonality as well as coinciding and unforeseen events. Therefore, it is no surprise that the assumptions of linear modeling are often ineffective in predicting cash needs.
A better and more intuitive methodology for forecasting cash is now available due to the advances in Artificial Neural Network (ANN) technologies that more closely replicate the way in which people think. Unlike linear methods, ANN methodology is ideally suited to identifying very complex cash use patterns and recursively adapting the pattern definitions over time.