| Download | - Will be available here on August 9, 2026
|
|---|
| DOI | Resolve DOI: https://doi.org/10.1016/B978-0-443-13293-3.00019-1 |
|---|
| Author | Search for: Cai, Xiatong; Search for: Mohammadian, Abdolmajid; Search for: Cobo, Hiedra Juan1; Search for: Shirkhani, Hamidreza1ORCID identifier: https://orcid.org/0000-0002-0893-652X; Search for: Imanian, Hanifeh; Search for: Payeur, Pierre |
|---|
| Affiliation | - National Research Council Canada. Construction
|
|---|
| Format | Text, Book Chapter |
|---|
| Abstract | The prediction of soil temperature under climate change plays an important role in understanding hydrological processes. Genetic programming can create a mathematical equation that can be used for predictions with very high efficiency due to its explicit analytical form. However, it is rarely used in soil temperature prediction, especially in extremely hot weather conditions. The fitness of multigene genetic programming (MGGP) in ordinary weather was found to be R² = 0.97, and R² = 0.83 in extremely hot weather. We compared the performance of single-gene genetic programming (SGGP) and multigene genetic programming (MGGP) with benchmark linear and AI models. Results show that the MGGP algorithm outperforms linear models and is comparable with some distance-based and tree-based benchmark AI models in both ordinary and extremely hot weather. MGGP underperformed the artificial neural network. Using only a polynomial equation rather than executing a complicated model with a large input dataset, MGGP shows good simplification in soil temperature prediction. |
|---|
| Publication date | 2024-08-09 |
|---|
| Publisher | Elsevier |
|---|
| In | |
|---|
| Language | English |
|---|
| Peer reviewed | Yes |
|---|
| Export citation | Export as RIS |
|---|
| Report a correction | Report a correction (opens in a new tab) |
|---|
| Record identifier | d779c85d-9968-49e8-bfea-a4cf1257a836 |
|---|
| Record created | 2023-11-02 |
|---|
| Record modified | 2024-08-23 |
|---|
