Comparative Analysis of MLR, ANN, and ANFIS Models for Prediction of Field Capacity and Permanent Wilting Point for Bafra Plain Soils

Taşan S., Demir Y.

COMMUNICATIONS IN SOIL SCIENCE AND PLANT ANALYSIS, vol.51, no.5, pp.604-621, 2020 (SCI-Expanded) identifier identifier

  • Publication Type: Article / Article
  • Volume: 51 Issue: 5
  • Publication Date: 2020
  • Doi Number: 10.1080/00103624.2020.1729374
  • Journal Indexes: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Academic Search Premier, PASCAL, Agricultural & Environmental Science Database, Aqualine, BIOSIS, CAB Abstracts, Chemical Abstracts Core, Chimica, Environment Index, Geobase, Pollution Abstracts, Veterinary Science Database
  • Page Numbers: pp.604-621
  • Keywords: Word, field capacity, permanent wilting point, multiple-linear regression, artificial neural network, adaptive neuro-fuzzy inference system, ARTIFICIAL NEURAL-NETWORK, PEDOTRANSFER FUNCTIONS, INFERENCE SYSTEM, HYDRAULIC-PROPERTIES, MULTIPLE-REGRESSION, WATER RETENTION, PLATEAU, MACHINE
  • Ondokuz Mayıs University Affiliated: Yes


Soil hydraulic parameters like moisture content at field capacity and permanent wilting point constitute significant input parameters of various biophysical models and agricultural practices (irrigation timing and amount of irrigation to be applied). In this study, the performance of three different methods (Multiple linear regression - MLR, Artificial Neural Network - ANN and Adaptive Neuro-Fuzzy Inference System - ANFIS) with different input parameters in prediction of field capacity and permanent wilting point from easily obtained soil characteristics were compared. Correlation analysis indicated that clay content, sand content, cation exchange capacity, CaCO3, and organic matter had significant correlations with FC and PWP (p < .01). Validation results revealed that the ANN model with the greatest R-2 and the lowest MAE and RMSE value exhibited better performance for prediction of FC and PWP than the MLR and ANFIS models. ANN model had R-2 = 0.83, MAE = 2.36% and RMSE = 3.30% for FC and R-2 = 0.81, MAE = 2.15%, RMSE = 2.89% for PWP in training dataset; R-2 = 0.80, MAE = 2.27%, RMSE = 3.12% for FC and R-2 = 0.83, MAE = 1.84%, RMSE = 2.40% for PWP in testing dataset. Also, Bayesian Regularization (BR) algorithm exhibited better performance for both FC and PWP than the other training algorithms.