Stochastic Environmental Research and Risk Assessment, Год журнала: 2024, Номер unknown
Опубликована: Сен. 28, 2024
Язык: Английский
Geophysical Research Letters, Год журнала: 2024, Номер 51(19)
Опубликована: Окт. 8, 2024
Abstract Black carbon (BC) is a highly persistent yet poorly understood component of forest soil reservoirs, while its inventory, distribution, and determining factors in soils on large geographic scale remain unclear. Here, we characterized BC across 68 Chinese sites using benzene polycarboxylic acid method developed machine learning (ML) models to predict interpret potential impacts organic matter (SOM) properties, physiochemical meteorological conditions, wildfire history, microbial diversity BC. Results revealed that SOM properties were the most critical predicting BC, complemented by negative impact mean annual temperature alkaline mineral composition. The superior prediction accuracy for with higher condensed aromaticity (more hexa‐ penta‐carboxylic monomers) likely results from simpler sources greater resistance transformation. This study introduces an effective ML model interpreting inventory better understand cycling.
Язык: Английский
Процитировано
1
Science Letters, Год журнала: 2024, Номер 12(1)
Опубликована: Фев. 12, 2024
This work aimed to study the modeling of organic pollution waters Déganobo Lake system by three models: Multiple Linear Regression model (MLR model), Mutilayer Perceptron (MLP model) and Regression/ hybrid (MLR/MLP model).In its implementation, chemical oxygen demand (COD) these waters, obtained from August 2021 July 2022, was used.Two approaches were done in case their COD MLP MLR/MLP model: static dynamic modeling.The results have highlighted low predictions MLR (36.2 %) models (6-8-1 for 7-3-1 modeling, both predicting less than 35% experimental values with high error (RMSE upper 1.30 relative 0.750).However, (MLR/6-3-1 MLR/7-3-1 modeling) well predicted around 99% very errors 0.0001 0.006 cases).So, most efficient predict waters.The accuracy this ecological again provided during study.
Язык: Английский
Процитировано
0
Science Letters, Год журнала: 2024, Номер 12(1), С. 1 - 9
Опубликована: Фев. 12, 2024
This work aimed to study the modeling of organic pollution waters Déganobo Lake system by three models: Multiple Linear Regression model (MLR model), Mutilayer Perceptron (MLP model) and Regression/ hybrid (MLR/MLP model). In its implementation, chemical oxygen demand (COD) these waters, obtained from August 2021 July 2022, was used. Two approaches were done in case their COD MLP MLR/MLP model: static dynamic modeling. The results have highlighted low predictions MLR (36.2 %) models (6-8-1 for 7-3-1 modeling, both predicting less than 35% experimental values with high error (RMSE upper 1.30 relative 0.750). However, (MLR/6-3-1 MLR/7-3-1 modeling) well predicted around 99% very errors 0.0001 0.006 cases). So, most efficient predict waters. accuracy this ecological again provided during study.
Язык: Английский
Процитировано
0
Science Letters, Год журнала: 2024, Номер 12
Опубликована: Фев. 13, 2024
This work aimed to study the modeling of organic pollution waters Déganobo Lake system by three models: Multiple Linear Regression model (MLR model), Mutilayer Perceptron (MLP model) and Regression/ hybrid (MLR/MLP model).In its implementation, chemical oxygen demand (COD) these waters, obtained from August 2021 July 2022, was used.Two approaches were done in case their COD MLP MLR/MLP model: static dynamic modeling.The results have highlighted low predictions MLR (36.2 %) models (6-8-1 for 7-3-1 modeling, both predicting less than 35% experimental values with high error (RMSE upper 1.30 relative 0.750).However, (MLR/6-3-1 MLR/7-3-1 modeling) well predicted around 99% very errors 0.0001 0.006 cases).So, most efficient predict waters.The accuracy this ecological again provided during study.
Язык: Английский
Процитировано
0
Stochastic Environmental Research and Risk Assessment, Год журнала: 2024, Номер unknown
Опубликована: Сен. 28, 2024
Язык: Английский
Процитировано
0