Table 1 Summary of literature review on load forecasting techniques.

Taylor et al. ¹⁰, Hahn et al. ¹¹

ARIMA, Linear Regression

Traditional load forecasting data

Simple and easy to visualize, but struggle with non-linear data and complex renewable patterns

Ineffective with non-linear data and complex patterns from renewable sources

Mystakidis et al.¹²

Linear Regression, Decision Trees

Energy prediction datasets

Models are interpretable and applicable to both regression and classification tasks

Prone to overfitting and less reliable for renewable variability

Gellert et al.¹³

Random Forest

Large energy datasets

Enhances prediction accuracy by combining multiple decision trees, reducing overfitting

Computationally intensive, especially when applied to large datasets, compared to simpler base models such as linear regression or decision trees

Chen and Guestrin¹⁵

XGBoost

Large, complex forecasting datasets

Superior speed and accuracy for complex forecasting tasks, efficient and scalable

Sensitive to hyperparameter settings and can be computationally demanding, especially when compared to simpler base models such as linear regression

Wang et al. ¹⁶

Support Vector Machines (SVM)

Energy forecasting data

Applicable to both linear and non-linear data for regression and classification tasks

Sensitive to parameter selection can be computationally heavy

Quinlan¹⁷

Decision Trees

Energy datasets

Straightforward and interpretable

Susceptible to overfitting, especially in complex datasets

Breiman¹⁸

Random Forest

Energy datasets

More stable and accurate predictions by aggregating multiple decision trees

Requires considerable computational resources

Cortes and Vapnik¹⁹

SVM

Diverse data structures in load forecasting

Robust in both classification and regression tasks

Computational complexity increases with dataset size

Al Arafat et al.²¹

Multi-Layer Perceptrons (MLPs)

Forecasting datasets

Captures intricate, non-linear interactions; adaptable across forecasting applications

Computationally intensive; requires careful tuning to avoid overfitting

Chen et al.²³

ML models with weather data integration

Energy datasets with external factors

Incorporating weather data improves forecasting accuracy

Requires accurate external data; forecasting accuracy depends on data quality

Li et al.²⁴

LSTM

Historical wind speed data (China)

Effective for short-term wind speed forecasting; captures long-term dependencies and handles time-series data well

Single-region data; lacks cross-regional validation