Time Series Forecasting Based on the Logistic Curve

This paper presents a class of models which are designed for forecasting the net sales of a product when the stock of that product is believed to be subject to a saturation level. The forecast function for the stock takes the form of a general modified exponential, a family which includes the logistic as a special case. However, framing the model in terms of the net increase in the product enables a link to be made between the traditional approach to forecasting based on non-linear trend curves and the approach based on ARIMA models.