Prediction of the Adsorption Behaviors of Radionuclides onto Bentonites Using a Machine Learning Method
This study builds a model to predict distribution coefficients (K<sub>d</sub>) using the random forest (RF) method and a machine learning model based on the Japan Atomic Energy Agency Sorption Database (JAEA-SDB). A database of ten input variables, including the distribution coefficient,...
| Main Authors: | , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
MDPI AG
2022-09-01
|
| Series: | Minerals |
| Subjects: | |
| Online Access: | https://www.mdpi.com/2075-163X/12/10/1207 |
| _version_ | 1797471095425597440 |
|---|---|
| author | Do-Hyeon Kim Jun-Yeop Lee |
| author_facet | Do-Hyeon Kim Jun-Yeop Lee |
| author_sort | Do-Hyeon Kim |
| collection | DOAJ |
| description | This study builds a model to predict distribution coefficients (K<sub>d</sub>) using the random forest (RF) method and a machine learning model based on the Japan Atomic Energy Agency Sorption Database (JAEA-SDB). A database of ten input variables, including the distribution coefficient, pH, initial radionuclide concentrations, solid–liquid ratio, ionic strength, oxidation number, cation exchange capacity, surface area, electronegativity, and ionic radius, was constructed and used for the RF model calculation. The calculation parameters employed in this work included two different hyperparameters, the number of decision trees and the maximum number of variables to divide each node, together with the random seeds inside the RF model. The coefficients of determination were derived with various combinations of hyperparameters and random seeds, and were employed to assess the RF model calculation result. Based on the results of the RF model, the distribution coefficients of 22 target nuclides (Am, Ac, Co, Cm, Cd, Cs, Cu, Na, Np, Ni, Nb, U, Sr, Sn, Pb, Pa, Pu, Po, I, Tc, Th, and Zr) were predicted successfully. Among the various input variables, pH was found to make the highest contribution to determining the distribution coefficient. The novelty of this study lies in the first application of the machine learning method for predicting the K<sub>d</sub> value of bentonites, using JAEA-SDB. This study has established a model for reliably predicting the distribution coefficient for various radionuclides that is intended for use in evaluating the K<sub>d</sub> value in arbitrary aqueous conditions. |
| first_indexed | 2024-03-09T19:44:41Z |
| format | Article |
| id | doaj.art-b685965520b64d5898dec7b6d28dbe92 |
| institution | Directory Open Access Journal |
| issn | 2075-163X |
| language | English |
| last_indexed | 2024-03-09T19:44:41Z |
| publishDate | 2022-09-01 |
| publisher | MDPI AG |
| record_format | Article |
| series | Minerals |
| spelling | doaj.art-b685965520b64d5898dec7b6d28dbe922023-11-24T01:28:45ZengMDPI AGMinerals2075-163X2022-09-011210120710.3390/min12101207Prediction of the Adsorption Behaviors of Radionuclides onto Bentonites Using a Machine Learning MethodDo-Hyeon Kim0Jun-Yeop Lee1School of Mechanical Engineering, Pusan National University, Busandaehak-ro 63beon-gil, Geumjeong-gu, Busan 46241, KoreaSchool of Mechanical Engineering, Pusan National University, Busandaehak-ro 63beon-gil, Geumjeong-gu, Busan 46241, KoreaThis study builds a model to predict distribution coefficients (K<sub>d</sub>) using the random forest (RF) method and a machine learning model based on the Japan Atomic Energy Agency Sorption Database (JAEA-SDB). A database of ten input variables, including the distribution coefficient, pH, initial radionuclide concentrations, solid–liquid ratio, ionic strength, oxidation number, cation exchange capacity, surface area, electronegativity, and ionic radius, was constructed and used for the RF model calculation. The calculation parameters employed in this work included two different hyperparameters, the number of decision trees and the maximum number of variables to divide each node, together with the random seeds inside the RF model. The coefficients of determination were derived with various combinations of hyperparameters and random seeds, and were employed to assess the RF model calculation result. Based on the results of the RF model, the distribution coefficients of 22 target nuclides (Am, Ac, Co, Cm, Cd, Cs, Cu, Na, Np, Ni, Nb, U, Sr, Sn, Pb, Pa, Pu, Po, I, Tc, Th, and Zr) were predicted successfully. Among the various input variables, pH was found to make the highest contribution to determining the distribution coefficient. The novelty of this study lies in the first application of the machine learning method for predicting the K<sub>d</sub> value of bentonites, using JAEA-SDB. This study has established a model for reliably predicting the distribution coefficient for various radionuclides that is intended for use in evaluating the K<sub>d</sub> value in arbitrary aqueous conditions.https://www.mdpi.com/2075-163X/12/10/1207adsorptionbentonitedistribution coefficientmachine learningrandom forest |
| spellingShingle | Do-Hyeon Kim Jun-Yeop Lee Prediction of the Adsorption Behaviors of Radionuclides onto Bentonites Using a Machine Learning Method Minerals adsorption bentonite distribution coefficient machine learning random forest |
| title | Prediction of the Adsorption Behaviors of Radionuclides onto Bentonites Using a Machine Learning Method |
| title_full | Prediction of the Adsorption Behaviors of Radionuclides onto Bentonites Using a Machine Learning Method |
| title_fullStr | Prediction of the Adsorption Behaviors of Radionuclides onto Bentonites Using a Machine Learning Method |
| title_full_unstemmed | Prediction of the Adsorption Behaviors of Radionuclides onto Bentonites Using a Machine Learning Method |
| title_short | Prediction of the Adsorption Behaviors of Radionuclides onto Bentonites Using a Machine Learning Method |
| title_sort | prediction of the adsorption behaviors of radionuclides onto bentonites using a machine learning method |
| topic | adsorption bentonite distribution coefficient machine learning random forest |
| url | https://www.mdpi.com/2075-163X/12/10/1207 |
| work_keys_str_mv | AT dohyeonkim predictionoftheadsorptionbehaviorsofradionuclidesontobentonitesusingamachinelearningmethod AT junyeoplee predictionoftheadsorptionbehaviorsofradionuclidesontobentonitesusingamachinelearningmethod |