Classification and Regression Machine Learning Models for Predicting Aerobic Ready and Inherent Biodegradation of Organic Chemicals in Water

Abstract

Machine learning (ML) is viewed as a promising tool for the prediction of aerobic biodegradation, one of the most important elimination pathways of organic chemicals from the environment. However, available models only have small datasets (<3200 records), make binary classification predictions, evaluate ready biodegradability, and do not incorporate experimental conditions (e.g., system setup and reaction time). This study addressed all these limitations by first compiling a large database of 12,750 records, considering both ready and inherent biodegradation under different conditions, and then developing regression and classification models using different chemical representations and ML algorithms. The best regression model (R 2 = 0.54 and root mean square error of 0.25) and classification model (the prediction accuracy from 85.1%) achieved very good performance. The model interpretation indicated that the models correctly captured the effects of chemical substructures, following the order of C�O > O�C−O > OH > CH 3 > halogen > branching > N > 6-member ring. The consideration of chemical speciation based on pK a and α notations did not affect the regression model performance but significantly improved the classification model performance (the accuracy increased to 87.6%). The models also showed large applicability domains and provided reasonable predictions for more than 98% of over 850,000 environmentally relevant chemicals in the Distributed Structure-Searchable Toxicity database. These robust, trustable models were finally made widely accessible through two free online predictors with graphical user interface.

Predicting THM Formation and Revealing Its Contributors in Drinking Water Treatment Using Machine Learning

“…Similar feature importance hierarchy was found for the LSBoost model (Figure S11). The mean SHAP value can explain the overall importance of features on model output; however, it cannot explain how each feature influences the formation of THMs. Figure b shows the impact of each feature on model prediction.…”

“…54,91 Positive or negative SHAP values for the features indicate whether the features have a positive or negative contribution for the THM formation, respectively. 89,90 For example, chlorine dose/DOC, reaction time, pH, and SUVA have low values (blue) on the left side and high values (red) on the right side, implying that increasing the value of these features will increase the THM formation. In contrast, for temperature, the majority of the high values are on the left side and the low values are on the right side, indicating a negative contribution for the THM formation.…”

Electricity-Driven Microbial Metabolism of Carbon and Nitrogen: A Waste-to-Resource Solution

“…Whereas DFT is frequently used to model the nature of active sites on electrocatalysts, 33,170 a robust modeling technique, e.g., machine learning, can be used for unveiling the complex effects of electric field on the microbial metabolism. 171 Recently, the ecological community dynamics was successfully predicted using community-level parameters during CO 2 and organic waste valorization. 172 Microbes can be engineered by using, for example, clustered regularly interspaced short palindromic repeats (CRISPR)-based genome editing, to improve its tolerance to toxic products.…”

“…Single-cell level and culture-independent methods may greatly help the understanding of the electrode-microbe interactions as well as syntrophic interactions of environmental samples. Whereas DFT is frequently used to model the nature of active sites on electrocatalysts, , a robust modeling technique, e.g., machine learning, can be used for unveiling the complex effects of electric field on the microbial metabolism . Recently, the ecological community dynamics was successfully predicted using community-level parameters during CO 2 and organic waste valorization .…”

Intelligent analysis of carbendazim in agricultural products based on a ZSHPC/MWCNT/SPE portable nanosensor combined with machine learning methods

“…Machine learning (ML) can mine hidden rules from large amounts of data and can be used for regression (prediction) or classification. 27 It has been widely used in material chemistry-related fields. 28 In recent years, Support Vector Machine (SVM), Least Square Support Vector Machine (LS-SVM) and Back Propagation-Artificial Neural Network (BP-ANN) have been widely used in the field of electrochemistry.…”