A Combination Short-Term Load Forecasting Based on Decision-Making

Firstly choosing five forecasting model that respectively giving a virtual load forecasting for the same case, then using the decision theory to sift models through odds matrix method, abandon the unsatisfactory models. Finally determine weightings according to each period of time minimum quadratic sum of residual error as objective function. Using the selected models for combination short-term load forecasting and comparing the result with the single model load forecasting results. It testified decision-making was referable and practical, the fitting accuracy of combination load forecasting was better than single one´s.