Machine Learning Based Prediction of Soil pH

Abstract

The significance of pH value of soil in assessing the soil fertility has been the motivation behind writing this research paper. As the pH of the soil is strongly dependent on the extent of minerals present in it, its pH can be estimated using the mineral data. The soil data belonging to the Kumaun part of Uttarakhand is used to design the prediction models. A machine learning techniaue SVM (support vector machine) has been implemented in MATLAB version R2021a. Three different prediction models with linear, quadratic and cubic kernels are established, and their prediction accuracy is compared.