|
At present, the electricity price forecasting method is time series, neural network, wavelet transform, the method is to point to forecast. This paper puts forward a kind of based on cloud model is a new method of short-term electricity price forecasting. This paper firstly introduces the concept and characteristics of cloud model, and gives the cloud model based on the price of electricity and load data discretization and concept jump process, obtained the electricity and the concept of load model. Through the great decision method for data collection soft division, the establishment of electricity price, load of Boolean type database, and then according to the given support and confidence soft thresholding based on cloud related knowledge mining algorithm, get time, load and electricity price between the association rules. Then with time, load conjunctions as a rule before a, the price of electricity as a rule, the establishment of a rule generator, according to dig out the rules for prediction. The method to get the prediction result is a series of uncertain discrete point set, set in each point can be predicted results provided to the user, the user can according to the experience and other information to the appropriate choice, also can put all the point expectations as a result provides to the user.
|