Comparative Analysis Between Feedforward Neural Network and CNN-LSTM Neural Network To Predict Household Electrical Energy Consumption
Conference
Tufail, S, Tariq, M, Batool, S et al. (2023). Comparative Analysis Between Feedforward Neural Network and CNN-LSTM Neural Network To Predict Household Electrical Energy Consumption
. 10.1109/ICECCME57830.2023.10253452
Tufail, S, Tariq, M, Batool, S et al. (2023). Comparative Analysis Between Feedforward Neural Network and CNN-LSTM Neural Network To Predict Household Electrical Energy Consumption
. 10.1109/ICECCME57830.2023.10253452
This paper compares the accuracy of energy prediction using Feedforward Neural Networks (FNN) with a hybrid Convolutional Neural Network- Long Short-Term Memory (CNN-LSTM) approach. The research builds two models, a FNN and a CNN-LSTM, and tests them on a large dataset of energy usage records. To extract local data and preserve long-range temporal relationships, the CNN-LSTM model draws on the best of convolutional and recurrent neural networks. Metrics like mean absolute error (MAE), root mean squared error (RMSE), and mean absolute percentage error (MAPE) are used to compare the two models' efficacy. The CNN-LSTM model shows superior accuracy and generalization capabilities over the FNN model when predicting energy usage. These results have the potential to improve energy management system accuracy and reliability, which in turn can increase building energy efficiency and sustainability.
