Research on Power User Intelligent Load Forecasting Method Based on Data State Drive

this paper proposed a variable input structure SVM prediction model based on state analysis. By identifying the key load characteristics in the load data and storing the load characteristics instead of the original load data, user load data can be realized more accurate and efficient. Based on the variable input structure SVM model of state analysis, the same state load is searched according to the results of state prediction in each period of the forecast day, and the same state historical load is used as the input factor of the model to predict. It effectively overcomes the interference of user power drift effect on load forecasting., and the forecasting accuracy is effectively improved.