Automated Machine Learning Based Crop Recommendation System

Abstract

Agriculture plays a crucial role in supplying food for the population, as well as contributing significantly to the country's Gross Domestic Product (GDP) in India. For farmers to achieve higher yields and profitability, it is essential to select a crop based on soil parameters. To simplify this process, a system for crop recommendation based on Machine Learning models has been developed. As a result, farmers may find it difficult to make informed decisions when using the Machine Learning approach since it is both time-consuming and exhaustive. Automated Machine Learning is being used to simplify and speed up the process. A machine learning algorithm uses an automatic selection of algorithms, features, and hyperparameters to make predictions, which can result in more accurate results. This study examines various Automated Machine Learning frameworks and compares the accuracy scores of different crop recommendation systems. H2O and AutoGluon achieved the highest accuracy score of 92.0%.