Comparison of Machine Learning Techniques to Estimate Increase in Crop Productivity

Crop productivity is one of the biggest challenge in India due to uncertain climate changes, water level lacking, inefficient modern tools usage and low cultivate land. Crop productivity can be increased by using machine learning in agriculture. Machine learning can help to estimate or predict that which climate and land may be better for the farming with different vegetation indices. A detailed analysis and application of machine learning classifies the initial stage towards adopting smart agriculture in India. This study aims to show that the machine learning method is most suitable for predicting agricultural production in a certain area, s tate, or season. The objective of this study is to enable data-driven agriculture. The most precise and quick technique to process the vast amount of agricultural data was discovered by comparing multiple regression models, including decision tree, linear regression, XGB regression, and random forest. The top 5 crops produced in India were discovered via data visualization techniques and amongst them the top 3 are selected based on area, state, and season to give elaborated analysis of crop distribution and yield in India with their accuracy.