Predicting soil phosphorus and studying the effect of texture on the prediction accuracy using machine learning combined with near-infrared spectroscopy

Rabie Reda; Taoufiq Saffaj; Salah Eddine Itqiq; Ilham Bouzida; Ouadi Saidi; Kaoutar Yaakoubi; Brahim Lakssir; Najib El Mernissi; El Mestafa El Hadrami

doi:10.1016/j.saa.2020.118736

Predicting soil phosphorus and studying the effect of texture on the prediction accuracy using machine learning combined with near-infrared spectroscopy

Spectrochim Acta A Mol Biomol Spectrosc. 2020 Dec 5:242:118736. doi: 10.1016/j.saa.2020.118736. Epub 2020 Jul 21.

Authors

Rabie Reda¹, Taoufiq Saffaj², Salah Eddine Itqiq³, Ilham Bouzida⁴, Ouadi Saidi⁴, Kaoutar Yaakoubi⁴, Brahim Lakssir⁴, Najib El Mernissi⁴, El Mestafa El Hadrami²

Affiliations

¹ University Sidi Mohamed Ben Abdellah, Faculty of Sciences and Techniques of Fez, Laboratory of Applied Organic Chemistry, Fez, Morocco; Moroccan Foundation for Advanced Science, Innovation & Research, MAScIR, Rabat, Morocco. Electronic address: reda.rabie@usmba.ac.ma.
² University Sidi Mohamed Ben Abdellah, Faculty of Sciences and Techniques of Fez, Laboratory of Applied Organic Chemistry, Fez, Morocco.
³ National School of Computer Science and Systems Analysis, Advanced Digital Enterprise Modeling and Information Retrieval Research Laboratory, Rabat, Morocco; Moroccan Foundation for Advanced Science, Innovation & Research, MAScIR, Rabat, Morocco.
⁴ Moroccan Foundation for Advanced Science, Innovation & Research, MAScIR, Rabat, Morocco.

PMID: 32759034
DOI: 10.1016/j.saa.2020.118736

Abstract

The estimation of soil phosphorus is essential for agricultural activity. The laboratory chemical analysis techniques are expensive and labor-intensive. In the last decade, near-infrared spectroscopy has been become used as an alternative for soil attributes analysis. It is a rapid technique, and inexpensive relatively. However, this technique requires a calibration step using different machine learning and chemometrics tools. This study aims to develop predictive models for total soil phosphorus and extractable phosphorus by the Olson method (P-Olson) using three regression methods, namely partial least squares (PLS), regression support vector machine (RSVM) and backward propagation neural network (BPNN), combined with a proposed variable selection algorithm (PARtest) and a genetic algorithm PLS (GA-PAS). Also, it aims to investigate the effect of the texture on the accuracy of the prediction. The results show that PARtest combined with PBNN outperform the other used algorithms with an R²_t = 0.86, RMSE_t = 1104 mg kg^-1, and RPD = 3.23 for the TP. For P-Olson the RSVM coupled with GA-PLS outperforms all other methods with an R²_t = 0.77, RMSE_t = 20.09 mg kg^-1, and RPD = 1.90. The use of hierarchical ascendant clustering (HAC) helps to reduce the heterogeneity of soil and helps to increase the quality of prediction. The obtained results show that the models for clayey and loamy soils yielded an excellent prediction quality with an R²t = 0.88, RMSEt = 857.33 mg kg^-1, and RPD = 4.10 using BPNN with PARtest for TP. Furthermore, an R² = 0.83 RMSE = 8.30 mg kg^-1, RPD = 11.00 3.11using RSVM with GA-PLS for P-Olson. Thus, the texture has a significant effect on the prediction accuracy.

Keywords: Machine learning algorithms; Near-infrared spectroscopy; Soil phosphorus; Texture effects; Variable selection algorithms.