Predictive modelling approach to data mining for forecasting electricity consumption

This paper presents an approach of data mining technique to predict electricity demand of a geographical region based on the meteorological conditions. The value prediction predictive data mining technique is implemented with the Artificial Neural Networks. The values of the factors such as temperature, humidity and public holiday on which electricity consumption depends and the daily consumption values constitute the data. Data mining operations are performed on this historical data to form a prediction model which is capable of predicting daily consumption provided the meteorological parameters. The steps of knowledge discovery of data process are implemented. The data is preprocessed and fed to neural network for training it. The trained neural network is used to predict the electricity demand for the given meteorological conditions.