Improving Visitor Traffic Forecasting in Brick-and-Mortar Retail Stores with Neural Networks

With accurate visitor traffic forecasts, retail businesses can optimise their staff schedule and stock distribution for increased profits. Existing models such as Linear Regression, ARIMA and other statistical methods are being employed for traffic predictions, but the accuracy of these models leaves room for improvement. Neural Network models show promising results on similar data. Therefore, a neural network architecture for retail traffic forecasting is presented and evaluated in this paper. With this model, a slight increase in forecast accuracy is achieved.