COMPUTATION EFFICIENT RELIABILITY OPTIMAL OUTCOMES FOR ELECTRICITY PRICES FORECASTING USING LSTM AND FCN ALGORITHM
Forecasting electricity prices is a critical issue in the energy sector since it aids in decisions about energy production, storage, and trade. In this paper, we offer a novel method for forecasting electricity prices that combines the Long Short-Term Memory (LSTM) and Fully Convolutional Network (FCN) deep learning models. Our strategy strives to increase computation effectiveness while assuring accurate and ideal forecasting results. To accomplish this, we first smooth the time series and eliminate outliers from the electricity price data during pre-processing. Then, using the pre-processed data, we train the LSTM and FCN models independently. The outputs of both models are then combined using a weighted average ensemble technique. Finally, we use the Mean Absolute Percentage Error (MAPE) and Root Mean Squared Error (RMSE) metrics to assess the performance of our methodology. Our experimental findings on the New York Independent System Operator (NYISO) dataset demonstrate that our suggested technique outperforms state-of-the-art methodologies in terms of forecasting performance, with lower MAPE and RMSE values.
Long Short-Term Memory (LSTM), Fully Convolutional Network (FCN), Mean Absolute Percentage Error (MAPE) and Root Mean Squared Error (RMSE), R squared