Comparison of LSTM and GRU for Predicting Market Clearing Price in Open Electricity Market

The Indian power exchange market is rapidly evolving. As a result of which consistent forecasting of Market Clearance Price (MCP) has become a very crucial task for trading electricity to any part of the country. An analysis of the fifteen-minute MCP forecasting of the Indian Energy Exchange (IEX) is done. Firstly, only a single data point is predicted using Long-Short-Term Memory (LSTM) based Recurrent Neural Network (RNN) and Gated Recurrent Unit (GRU) based RNN. Furthermore, twenty data points which represent five hours ahead forecasting are done using both models. The comparative analysis between the experimental results and the real data from IEX proves the effectiveness of the developed models that GRU is better when it comes to speed and memory, but in terms of accuracy LSTM has got a slight edge over GRU.