Eclipse calculates demand forecasting on a regular basis. With the advanced forecast methods, you may not know which of the advanced methods historically is closest to the actual sales. The traditional Eclipse forecasting is a flat demand-per-day calculation. The Advanced Demand Forecasting Analysis window displays these best fit selections.
To know whether or not you are using the best forecast method for your products, Eclipse runs on stock products you select, such as specific buy line, and allows you to compare your previous demand forecast to what the system calculates as the best fit. The system runs as if the product is set to standard method for each method individually and without relying on information from the other calculations.
Eclipse uses a Mean Squared Error (MSE) calculation to determine the best fit demand forecast for your product(s).
This MSE compares the past 52 weeks of demand forecast to actual sales. If the system does not find 52 weeks worth of data, the system uses as many weeks as are available. For each available week, the lowest calculated MSE is the best fit.
Use the Auto Set Best Fit option on the Advanced Demand Forecasting window to automatically use the best fit calculation for the selected products. This means if Eclipse finds that a different forecast method is a better fit for a specific product or product line, then the system overrides that forecast method over the one that has been selected. These settings are at the branch level. Review your product settings carefully before using this option.
Run the demand for each week back for 52 weeks for each demand forecast method to get a full year's worth of information for comparing the values. The system then performs the same check for the three advanced methods. The system can then display which of the six methods is a best fit for your demand forecasting.
Note: Eclipse considers a standard week for calculations is Sunday to Saturday.