Abstract
Most methods of projecting past sales data into the future require this calculation of moving averages so as to smooth the data and obtain a trend. This can be cumbersome to handle on a computer, particularly if a large number of products are being forecast as a routine, and the amount of storage space required is excessive. The methods propounded here all make use of exponentially weighted moving averages, which are very readily calculated, require a minimum of storage space, and have the advantage in that they put declining weight on older data.

This publication has 0 references indexed in Scilit: