Balancing accuracy and efficiency in multi-step building-level electric load forecasting: Deep learning vs. classical time-series models

Chemical Engineering

• Deep Learning Method 100%
• Bayesian Optimization 100%
• Long Short-Term Memory 80%
• Long Short-Term Memory Network 80%
• Recurrent Neural Network 20%

Engineering

• Electric Load Forecasting 100%
• Consumption Pattern 25%
• Smart Energy Management 25%
• Load Forecasting 25%
• Input Uncertainty 25%