Comparative analysis of seven machine learning algorithms and five empirical models to estimate soil thermal conductivity

•The accuracy of 7 ML algorithms and 5 thermal conductivity models was evaluated.•The average RMSE values of ML algorithms were 66%∼82% of the empirical model values.•Soil moisture and bulk density together had > 80% of the feature importance value. Soil thermal conductivity (λ) is an important t...

Celý popis

Uloženo v:
Podrobná bibliografie
Vydáno v:Agricultural and forest meteorology Ročník 323; s. 109080
Hlavní autoři: Zhao, Tianyue, Liu, Shuchao, Xu, Jia, He, Hailong, Wang, Dong, Horton, Robert, Liu, Gang
Médium: Journal Article
Jazyk:angličtina
Vydáno: Elsevier B.V 15.08.2022
Témata:
decision support systems
Empirical model
empirical models
energy balance
Feature importance ranking
forests
Machine learning
meteorology
normal distribution
Random forest
regression analysis
soil density
Soil thermal conductivity
soil water content
thermal conductivity
Thermal conductivity model
NN
LR
GP
Random forest
GBDT
KNN
Empirical model
Feature importance ranking
DT
RF
Soil thermal conductivity
Thermal conductivity model
Machine learning
ML
ISSN:0168-1923, 1873-2240
On-line přístup:Získat plný text
Tagy: Přidat tag
Žádné tagy, Buďte první, kdo vytvoří štítek k tomuto záznamu!
Popis
Shrnutí:•The accuracy of 7 ML algorithms and 5 thermal conductivity models was evaluated.•The average RMSE values of ML algorithms were 66%∼82% of the empirical model values.•Soil moisture and bulk density together had > 80% of the feature importance value. Soil thermal conductivity (λ) is an important thermal property that is crucial for surface energy balance and water balance studies. 1602 measured soil thermal conductivity values representing 189 soils were used to evaluate five empirical models (i.e., de Vries (1963) model (de Vries 1963), Campbell (1985) model (Campbell1985), Johansen (1975) model (Johansen 1975), Côté and Konrad (2005) model (Côté and Konrad 2005), and Lu et al. (2007) model (Lu 2007)) and seven machine learning (ML) algorithms (i.e., Decision Tree (DT), Random Forest (RF), Gradient Boosting Decision Tree (GBDT), Linear Regression (LR), K-Nearest Neighbors (KNN), Neural Network (NN), and Gaussian Process (GP)) to estimate λ. Our results demonstrated that the average root mean squared error (RMSE) values of ML were 66% and 82% of the empirical model values on validation and test sets respectively. The three best ML algorithms (GBDT, NN, RF) performed significantly better than the three best empirical models (Lu 2007, Côté and Konrad 2005, Johansen 1975): 0.183 < RMSE < 0.259 (W m−1 K−1) for ML algorithms and 0.293 < RMSE < 0.320 (W m−1 K−1) for empirical models. For ML, we recommend the GBDT, NN and RF algorithms. For empirical models, we recommend to use three normalized models (Lu 2007, Côté and Konrad 2005, Johansen 1975) over the physically-based model (DV1963) and the regression model (CG1985). The feature importance rankings performed by the RF and GBDT algorithms show that soil moisture content and soil bulk density are the most critical factors affecting λ. Soil moisture content and soil bulk density together account for more than 80% of the influence importance value of λ. RF gives more consistent feature importance ranking results than GBDT, therefore, we recommend the use of RF for selecting features.
Bibliografie:ObjectType-Article-1
SourceType-Scholarly Journals-1
ObjectType-Feature-2
content type line 23
ISSN:0168-1923
1873-2240
DOI:10.1016/j.agrformet.2022.109080