Artificial Neural Network (ANN) based Soil Electrical Conductivity (SEC) prediction

In this study, the ANN-based model has been developed to predict the SEC. SEC is an important indicator of soil degradation. If the SEC is high then vegetation growth will be decreased. The reason for the high SEC is the presence of soluble salts in the soil system. Environmental parameters were taken as covariates in developing the ANN model. Various covariates used in the study are, NDVI (Normalized Differential Vegetation Index), Rainfall, DEM (Digital Elevation Model) and slope of the study region. SEC data of the study region was obtained from Soil Health Card (SHC) for the year 2014 and was incorporated in training the model. The SHC data was divided in different ratios like 60:40, 70:30, and 80:20 for training and testing the model. The SCORPAN model has been considered as the base model for the development of the ANN based prediction model. The highest coefficient of determination (r square) between actual and predicted SEC was 0.63 and MSE (Mean Square Error) was 0.04 at 80:20 training, testing the samples.