PREDICTING ELECTRICITY LOAD FOR LIGHT ENERGY CONSUMPTION IN A HOUSE USING MACHINE LEARNING ALGORITHM

Predicting energy consumption is an area has become an important field of research. The purpose of this research work is to find the relation between the usage of light energy consumption with outside temperature and humidity. In this work, time series modeling data of electric light energy consumption for a single house is considered from January 2016 to May 2016. Artificial neural network has various time series models for predicting time series data such as NAR and NARX. In this work, the error rate and accuracy value of NAR and NARX time series models are compared. NARX model provides better results when compared to NAR model. Artificial neural network has three different training algorithms like Levenberg-Marquardt, Bayesian Regularization and scaled Conjugate Gradient algorithms. These algorithms have different levels of accuracy value and error rate for the observed data and predicted data. Bayesian Regularization training algorithm provides the best accuracy and error rate for light energy consumption.