Modelling of the adsorption of urea herbicides by tropical soils with an Adaptive‐Neural‐based Fuzzy Inference System

Abstract

Sorption of pesticides by soils holds a major consequence for their fate in the environment. As such, sorption coefficient (K d /K oc), which is derived from laboratory or field experiments is a fundamental parameter used in almost all screening tools to evaluate the fate or mobility of these compounds. The value of this coefficient is controlled by many soil and solute specific properties, as well as environmental variables. Soft computing techniques such as Adaptive Neuro-Fuzzy Inference System (ANFIS) have been successfully used to predict the equilibrium partitioning of many compounds in various engineered systems. Application of these techniques to natural systems such as soils is however lacking. Here, we present the use of ANFIS in predicting the sorption per unit mass of soil, Q e , used in the calculation of K d /K oc of compounds in soils. In a previous study, we collected data associated to the adsorption of five phenylurea herbicides in 18 tropical soils. Here, we analysed such data and based on established correlations, nine variables were selected as potential input vectors (i.e., six soil properties, two herbicides molecular descriptors and initial solute concentrations). A total of 255 ANFIS models of one to eight input vectors were elaborated under 10-fold cross-validation. Multiple linear regression (MLR) models were similarly developed, and compared with the ANFIS in terms of mean absolute error (MAE), root-mean-square error (RMSE) and coefficient of determination (R 2). The best ANFIS model (M94) has an MAE test , RMSE test and R 2 test of 3.43 ± 0.43, 4.94 ± 0.80 and 0.95 ± 0.01, respectively, whereas the best MLR model (M13) returned an

Adaptive neuro-fuzzy inference system (ANFIS) and multiple linear regression (MLR) modelling of Cu, Cd, and Pb adsorption onto tropical soils

“…1and 2, For 1 L (V) of a specified initial metal ion concentration and 1 kg of soil (Ms), K d can be calculated from the values of Q e (experimental or predicted) by Eq. 4 (Agbaogun et al 2021).…”

“…where f i is output and p i , q i , and r i are the consequent parameters of the ith rule (Agbaogun et al 2021). A i and B i are linguistic labels whose membership function parameters…”

“…Recently, soft computing methods such as artificial neural network (ANN), fuzzy logic (FL)-based techniques (i.e. fuzzy inference systems, FIS), genetic algorithms (GA), and several connectionist systems such as adapted neuro-fuzzy inference system (ANFIS) are increasingly being recognised as accurate learning schemes for modelling complex phenomena in different aspects of engineering, physical, and natural sciences (Agbaogun et al 2021). These techniques have the capability to take care of uncertainties that often accumulate in traditional mathematical and statistical techniques (Kebria et al 2018).…”

“…This inference system has learning capability to approximate nonlinear functions (Aqil et al 2007) and particularly gaining popularity in dealing with ill-defined and uncertain domains; hence, it is considered a "universal estimator". Further, it is well-known for its ability to produce systems with good interpretability-accuracy trade-off by combining the advantages of ANN and FIS in a single framework and equally reduces the drawbacks of the two (Rahimzadeh et al 2016;Agbaogun et al 2021).…”

“…These include, but not limited to, Qasaimeh et al (2012), Amiri et al (2013), Ghaedi et al (2013Ghaedi et al ( , 2014, Tanhaei et al (2017), Baziar et al (2017), Dolatabadi et al (2018), Lashkenari et al (2018), Javadian et al (2018), andSigmund et al (2020). Most recently too, Agbaogun et al (2021) successfully used ANFIS to model the adsorption of phenylurea herbicides by soils. So far, our literature search has revealed that ANFIS has never been used in modelling the adsorption of heavy metal ions onto natural soils.…”