Appliance Energy Prediction using Time Series Forecasting: A Comparative Analysis of Different Machine Learning Algorithms

Energy prediction is critical for the overall economic development of a nation. Over the years, the implementation of several new procedures and techniques took place to predict the consumption of energy. Machine learning and statistical approach has been used and is very useful in the prediction. However, with the complexity of multiple devices in action, a varied number of parameters has to be taken care of the increasing amount of data. Hence, the demand for more advanced techniques and algorithms were required. In this paper, the focus is on time series forecasting. With the advent of deep learning and more advanced algorithms, time series forecasting came into the shadows. According to the problem Ensemble methods, Sarimax and LSTM used for the prediction and during the analysis, it was observed that time series forecasting, provided the best output.