Journal of Mines, Metals and Fuels

Vaidehi Bhaskara ¹, Ramesh K T ¹, Sayan Chakraborty ²

Affiliations
1 Department of Industrial Engineering and Management, B.M.S College of Engineering, Bengaluru, IN
2 Department of Operations and IT, ICFAI Business School, Hyderabad, IN

Abstract

More than half of the Indian population depends on agriculture as a source of livelihood. But, India’s marginal farmers especially, earn meagre amounts from their harvested yields. This may be attributed partly due to relatively smaller land holdings, and partly due to minimal access to resources that aid with informative price forecasts. In order to alleviate the stress caused by the lack of sound financial planning, this research proposes the utilization of machine learning to predict commodity prices. The solution obtained through such a model would assist farmers in predicting the price and associated estimates can be made with respect to yield, sowing patterns and suitable recommendations for sales. The solution developed in this research is a result of a thorough exploration of the literature in this domain, identification of verified secondary sources for data collection, and proposes a methodology to design a machine learning model that predicts prices for seasonal cash crops specific to the markets of Karnataka. Cotton has been used as the crop of focus in this study. ARIMA and Bayesian ridge regression have been used for predictive analytics, and the results obtained indicate a high correlation between the predicted and actual.

Keywords

Agriculture, Cash crop, Price prediction, Machine learning, Regression.

Full Text