Batch Processing & Reporting › CA PMA Chargeback Processing › Batch Forecasting (CAKRFCSB) › Forecast Methods Summary

Forecast Methods Summary

CA PMA Chargeback allows for three types of forecast methods: Manual, LINE and SMOOTH. All of these methods extract actual data from the CCCTAB table and either copies the data, or uses it to forecast data to the CCCMOD table. Once a CCCMOD table is created, the Period functions online facility can be used to perform what if analysis on the charge and unit consumption values.

Manual

To perform a manual forecast, specify COPY for the ALGORITHM control statement. Once a copy of the selected records is made, you can adjust the values displayed on the various online forecasting panels.

LINE

The Line algorithm computes the weighted moving average for the periods requested based upon historical data. It computes an average of the range of periods specified by the START_PERIOD and END_PERIOD control statements, giving each period a weight. The more recent the period, the greater the weight it is given. The average computed is then used as a forecast period.

If the value specified for the NUMBER_TO_FORECAST control statement is greater than one, each period forecasted is used to generate subsequent forecast periods. You can forecast up to six periods.

SMOOTH

The Smooth algorithm is an exponential method of forecasting with trending. This method bases its forecast on recent consumption values rather than on remote ones. Two smoothing constants, ALPHA and DELTA are used to compute the forecast.

ALPHA is used to keep up with the changes in demand. It can be any value between 0 and 1. If the charge and unit consumption of the records of the CCCTAB vary slightly from one period to the next, a small value of alpha (<.4 ) should be used. If the values in the CCCTAB are increasing or decreasing quickly, a larger value of alpha (>.5) should be used.

Often over a period of time, the exponential forecast slightly differs from that of the actual occurrence. This is due to the upward or downward shifts in data. DELTA, a smoothing constant is part of a trend adjustment used to counter the affects of these shifts. It can be any value between 0 and 1.

Note: Forecast algorithms use only actual data from the CCCTAB to perform the forecast. Minimum and maximum charge values are not applied.