Soil Testing Prediction System

Abstract

Abstract: Soil Testing Prediction is aimed to predict the soil functional properties (Calcium, Phosphorus, pH, Sand and Soil Organic Carbon) of a soil sample. Soil Testing Prediction finds its application in the field of agriculture, farming and research. It can help in economic crop management and better yield of crops. We have worked to find out how traditional soil testing methods can be replaced with modern Machine Learning techniques that can result in more economic , time efficient methods with no or little to no adverse effects on the environment. It trains to reduce the technical expertise required at users end, and aims to bring the labs to the user instead of taking the user to the lab. Keywords: 1) Linear Regression: Linear Regression is one of the many statistical techniques that has been adopted by the Machine Learning Society. It is a supervised learning technique i.e It works on labeled data. It assumes that there exists a linear relationship between the dependents and predictor. Although preemptive, the technique proves out to be at par with many ML techniques. 2) Feature Selection: Feature Selection is the method of extracting out the useful features from the set of all available features, that can help us to predict the required targets . It aims to reduce the error in the prediction made by the model. 3) Soil Functional Properties: Functional properties are the properties that define the behavior of the soil and its response to the environment and surroundings. 4) Mehlich-3 Extraction Techniques: It is a chemical method to predict the value over a range of elements. It is a weak acid soil extraction procedure. The extract is composed of 0.2 M glacial acetic acid, 0.25 M ammonium nitrate, 0.015 M ammonium fluoride, 0.013 M nitric acid, and 0.001 M ethylene diamine tetraacetic acid.