Electricity load forecast in the field of energy is one of the very important research topics. It is of very important application value to guarantee the safety of the system, energy savings and production efficiency maximization.The main research of the electricity load forecast is based on data classification methods and clustering analysis of data mining , and in the original basis of the three database model put forward on electricity load forecasting new models based on data mining techniques, and proposed new electricity load forecasting methods based on fuzzy inference and chaos time series digging. Combing subtraction clustering and fuzzy inference systems to create ANFIS(adaptive neural fuzzy inference system) to forecast short term electricity load data effectively, and Combing Chaos Theory and RBF Neural Network to create a load forecasting model for forecasting middle term electricity load reasonably and effectivly.This thesis addresses the current shortage of electricity load forecast studies and difficulties, based on data mining techniques focus on the issues of electricity load forecasting models and methods of research ,which is how to choose the inputs of load forecasting models; how to predeal with the inputs of fuzzy neural network electricity load forecasting methods ,and Chaos time series application. Through enormous data storage, data mining and decision support information, it can effectively overcome the limited nature of the data, the integrity of the data, and the effect of complexity of the factors affecting the outcome of the forecast. It has a unique advantage and can be of economic value.
|