Supervised Machine Learning Technique to Predict Soil Health

Abstract

In India agriculture is one of the major field which is been ignored by technical touch. Applying Artificial intelligence derivatives like Machine learning and Deep Learning to agricultural practices helps in maximum production of crops and maintains field’s soil health. Health of agricultural field mainly involves maintaining of soil nutrient like chemical and physical property of soil by properly channelizing the supplements. If soil health is managed scientifically, it progressively helps in high production of yield and long life of cultivation land.

Ontology is built on the soil data collected from soil testing centers. Ontology is built in a way which exhibits the knowledge and relationship between soil and its chemical nutrient. The knowledge base is then considered to relate the nutrient and the soil type.

To manage soil health and classify them to healthy and unhealthy class machine learning comes with handy and best in class algorithms. In this study evident algorithms of machine learning are used to classify the soil efficiently to two classes healthy and unhealthy.

Algorithms like logistic regression, Decision tree, Random tree classifier, Support Vector Machine and XGBoost was used to classify the data and their algorithmic efficiency was increased by hyper parameter tuning by using different techniques.