At present, Price Forecasting has many methods such as time-series, neural network, wavelet transform and so on. These are all the point forecasting to the price, This paper presents a method of the short-term price forecasting Based on cloud model. At first, this paper introduces the concept and the characteristics of the cloud model, and then presents the process of data discrete and concept zooming based on the cloud model, gives the concept models of the price and the load. This paper uses the Great Determination Law to distinct the data, create the Boolean database of price and load. According to the given degree of support and confidence of soft domain values, using the Cloud Association Rules to get the association rules. Then we take time and load as the rules of the former, price as the rules of the later, and establish rules generator, use the mining rules to predict. The result is a series of uncertainty discrete points. Each point can be provided to users as a result. Users can select appropriate results according to experience and other information. It can also provide the expectation for users.
|