Integration of machine learning-based prediction for enhanced Model's generalization: Application in photocatalytic polishing of palm oil mill effluent (POME). (December 2020)

• Record Type: Journal Article
• Authors: Ng, Kim Hoong; Gan, Y.S.; Cheng, Chin Kui; Liu, Kun-Hong; Liong, Sze-Teng
• Is Part Of: Environmental pollution. Volume 267(2020)
• Publication Date: 2020-12
• Subjects: Photocatalytic treatment -- Palm oil mill effluent -- Machine learning predictive models -- Generalization enhancement; Pollution -- Periodicals; Pollution -- Environmental aspects -- Periodicals; Environmental Pollution -- Periodicals; Pollution -- Périodiques; Pollution -- Aspect de l'environnement -- Périodiques; Pollution -- Effets physiologiques -- Périodiques; Pollution; Pollution -- Environmental aspects; Periodicals; Electronic journals; 363.73
• DOI: 10.1016/j.envpol.2020.115500
• Languages: English
• ISSNs: 0269-7491

Abstract

Abstract: In predicting palm oil mill effluent (POME) degradation efficiency, previous developed quadratic model quantitatively evaluated the effects of O 2 flowrate, T i O 2 loadings and initial concentration of POME in labscale photocatalytic system, which however suffered from low generalization due to the overfitting behaviour. Evidently, high RMSE (131.61) and low R2 (−630.49) obtained indicates its insufficiency in describing POME degradation at unseen factor ranges, hence verified the fact of poor generalization. To overcome this issue, several models were developed via machine learning-assisted techniques, namely Gaussian Process Regression (GPR), Linear Regression (LR), Decision Tree (DT), Supported Vector Machine (SVM) and Regression Tree Ensemble (RTE), subsequently being assessed systematically. To achieve high generalization, all models were subjected to 'train-all-test-all' strategy, 5-fold and 10-fold cross validation. Specifically, GPR model was furnished with high accuracy in 'train-all-test-all' strategy, judging from its low RMSE (1.0394) and high R2 (0.9962), which however menaced by the risk of overfitting. In contrast, despite relatively poorer RMSE and R2 (1.7964 and 0.9886) obtained in 5-fold cross validation, GPR model was rendered with highest generalization, while sufficiently preserving its accuracy in development process. Besides, SVM and RTE models were also demonstrated promising R2 (0.9372 and 0.9208), which however shadowed by their high RMSEs (4.2174 and 4.7366). Furthermore, the extraordinary generalization of GPR model was coincidentally verified in 10-fold cross validation. The lowest RMSE (2.1624) and highest R2 (0.9835) obtained with feature number of 36 asserted its sufficiency in both generalization and accuracy prospect. Other models were all rendered with slight lower R2 (> 0.9), plausibly due to the higher RMSE (> 4.0). According to GPR model, optimized POME degradation (52.52%) can be obtained at 70 mL/min of O2, 70.0 g/L of TiO2 and 250 ppm of POME concentration, with only ∼ 3% error as compared to the actual data. Graphical abstract: Image 1 Highlights: Machine-learning approaches were integrated to avoid overfitting circumstances. Models were assessed by train-all-test-all, 5-fold and 10-fold cross validations. Gaussian Process Regression model exhibited the best generalization and adequacy. Abstract : The Gaussian Process Regression (GPR) model, from machine learning algorithm, exhibited high generalization and adequacy in correlating the photocatalytic degradation of POME with its operating conditions.