| 摘要 |
A method and system predict future revenue of an account or a customer associated therewith for a specific time period based on past revenue. An exemplary customer revenue prediction system (100) includes a revenue prediction process (102) having access to a customer database (104), weight information (106) and adjustment rules (110). The technique uses historical revenue data of a predetermined number of time periods before the specific time period, to calculate the prediction revenue for the specific time perio d. Different weights are assigned to the revenue data of the predetermined numb er of time periods, wherein the weight is selected based on the recentness of each of the predetermined number of time periods relative to the specific ti me period. For instance, the revenue of a month closer to a month to be predict ed is given more weight than older months. The weight for each time periods may be determined empirically, such as by regression. Prediction revenue of the specific time period is determined based on the historical revenue data and the weight of each of the predetermined number of time periods. The predicti on revenue may be further adjusted to reflect growth rate. |
