Optimizing Solar Radiation Prediction with ANN and Explainable AI-Based Feature Selection
Reliable and accurate solar radiation (SR) prediction is crucial for renewable energy development amid a growing energy crisis. Machine learning (ML) models are increasingly recognized for their ability to provide accurate and efficient solutions to SR prediction challenges. This paper presents an A...
Saved in:
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2025-06-01
|
| Series: | Technologies |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2227-7080/13/7/263 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1839615122763415552 |
|---|---|
| author | Ibrahim Al-Shourbaji Abdalla Alameen |
| author_facet | Ibrahim Al-Shourbaji Abdalla Alameen |
| author_sort | Ibrahim Al-Shourbaji |
| collection | DOAJ |
| description | Reliable and accurate solar radiation (SR) prediction is crucial for renewable energy development amid a growing energy crisis. Machine learning (ML) models are increasingly recognized for their ability to provide accurate and efficient solutions to SR prediction challenges. This paper presents an Artificial Neural Network (ANN) model optimized using feature selection techniques based on Explainable AI (XAI) methods to enhance SR prediction performance. The developed ANN model is evaluated using a publicly available SR dataset, and its prediction performance is compared with five other ML models. The results indicate that the ANN model surpasses the other models, confirming its effectiveness for SR prediction. Two XAI techniques, LIME and SHAP, are then used to explain the best-performing ANN model and reduce its complexity by selecting the most significant features. The findings show that prediction performance is improved after applying the XAI methods, achieving a lower MAE of 0.0024, an RMSE of 0.0111, a MAPE of 0.4016, an RMSER of 0.0393, a higher <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><msup><mi>R</mi><mn>2</mn></msup></semantics></math></inline-formula> score of 0.9980, and a PC of 0.9966. This study demonstrates the significant potential of XAI-driven feature selection to create more efficient and accurate ANN models for SR prediction. |
| format | Article |
| id | doaj-art-022c0df82ec64b48ab7e6ec4f15dbed1 |
| institution | Matheson Library |
| issn | 2227-7080 |
| language | English |
| publishDate | 2025-06-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Technologies |
| spelling | doaj-art-022c0df82ec64b48ab7e6ec4f15dbed12025-07-25T13:37:15ZengMDPI AGTechnologies2227-70802025-06-0113726310.3390/technologies13070263Optimizing Solar Radiation Prediction with ANN and Explainable AI-Based Feature SelectionIbrahim Al-Shourbaji0Abdalla Alameen1Department of Electrical and Electronics Engineering, Jazan University, Jazan 45142, Saudi ArabiaDepartment of Computer Engineering and Information, Prince Sattam bin Abdulazizd University, Wadi Alddawasir 11991, Saudi ArabiaReliable and accurate solar radiation (SR) prediction is crucial for renewable energy development amid a growing energy crisis. Machine learning (ML) models are increasingly recognized for their ability to provide accurate and efficient solutions to SR prediction challenges. This paper presents an Artificial Neural Network (ANN) model optimized using feature selection techniques based on Explainable AI (XAI) methods to enhance SR prediction performance. The developed ANN model is evaluated using a publicly available SR dataset, and its prediction performance is compared with five other ML models. The results indicate that the ANN model surpasses the other models, confirming its effectiveness for SR prediction. Two XAI techniques, LIME and SHAP, are then used to explain the best-performing ANN model and reduce its complexity by selecting the most significant features. The findings show that prediction performance is improved after applying the XAI methods, achieving a lower MAE of 0.0024, an RMSE of 0.0111, a MAPE of 0.4016, an RMSER of 0.0393, a higher <inline-formula><math xmlns="http://www.w3.org/1998/Math/MathML" display="inline"><semantics><msup><mi>R</mi><mn>2</mn></msup></semantics></math></inline-formula> score of 0.9980, and a PC of 0.9966. This study demonstrates the significant potential of XAI-driven feature selection to create more efficient and accurate ANN models for SR prediction.https://www.mdpi.com/2227-7080/13/7/263Artificial Neural NetworkExplainable AIsolar radiation predictionoptimization |
| spellingShingle | Ibrahim Al-Shourbaji Abdalla Alameen Optimizing Solar Radiation Prediction with ANN and Explainable AI-Based Feature Selection Technologies Artificial Neural Network Explainable AI solar radiation prediction optimization |
| title | Optimizing Solar Radiation Prediction with ANN and Explainable AI-Based Feature Selection |
| title_full | Optimizing Solar Radiation Prediction with ANN and Explainable AI-Based Feature Selection |
| title_fullStr | Optimizing Solar Radiation Prediction with ANN and Explainable AI-Based Feature Selection |
| title_full_unstemmed | Optimizing Solar Radiation Prediction with ANN and Explainable AI-Based Feature Selection |
| title_short | Optimizing Solar Radiation Prediction with ANN and Explainable AI-Based Feature Selection |
| title_sort | optimizing solar radiation prediction with ann and explainable ai based feature selection |
| topic | Artificial Neural Network Explainable AI solar radiation prediction optimization |
| url | https://www.mdpi.com/2227-7080/13/7/263 |
| work_keys_str_mv | AT ibrahimalshourbaji optimizingsolarradiationpredictionwithannandexplainableaibasedfeatureselection AT abdallaalameen optimizingsolarradiationpredictionwithannandexplainableaibasedfeatureselection |