Paddy Yield Forecasting using Regression Techniques

Abstract

India is one of the well-known countries whose socioeconomic position is heavily influenced by agriculture. Plant and livestock cultivation are the most important practices in agriculture. The cultivation of food surpluses enables people to live in cities. Agriculture employs over 70 percent of Tamil Nadu's population. It is a key source of income in Tamil Nadu. Paddy cultivation is significant as it is the main food for the majority of the Indians. The dataset used for this study includes various varieties of paddy grown in the Madurai district of Tamil Nadu. The dataset includes average paddy crop yield, Madurai taluks, soil type, soil pH, nitrogen, temperature, duration, and rainfall. This paper compares different regression algorithms and suggests an optimal algorithm that is used in a ML (Machine Learning) model that forecasts paddy yield. The machine learning regression techniques are multiple linear regression, lasso regression, and ridge regression. This paper concludes by discussing future work topics and the suitable algorithm to be employed in the model for yield prediction of a paddy crop.