Factors affecting agriculture and estimation of crop yield using supervised learning algorithms

Abstract

India is a well-renowned agricultural nation in the world. It is among the prominent countries in the world, famous for surplus production of crops. Agriculture and the related activities are the main source of livelihood and income for more than 70% of the population of India. It plays a vital role in the economic growth of the country by contributing to around 14 to 15 % of the Gross Domestic Product (GDP). India has a strong backbone in the Indian economy due to the significant contributions made by the farmers. Modern-day advancements have left an indelible imprint on the agriculture pattern of the country. It is due to the negligence of people which results in the deterioration of the harvest yield. In this work an Artificial Intelligent approach on harvest yield is performed to assess various factors affecting the yield of the crops in Tamil-Nadu state between the years 2000 to 2015, which primarily include: rainfall patterns to recommend different crops based on Nitrogen, Phosphorus, Potassium, Temperature, Humidity and pH values of soil. A comparative analysis is performed on various supervised learning algorithms. XGBoost model for suggesting crop harvest based on parameters affecting soil quality is deployed, providing an overall test accuracy of 99.318%, outperformed other discussed techniques like DT, Gaussian Naive Bayes (GNB), SVM and other similar types of supervised models. The comprehensive approach developed to estimate crop sustainability using supervised algorithms helps in increasing farm yield, reduces manual work, time spent on various agriculture activities and recommends crop based on given soil parameters.