Machine learning regression model for material synthesis prices prediction in agriculture

Abstract

Agriculture is the back bone of India, one of the major sources for some notable population from past decade, slowly the agriculture is facing downward growth. Due to less rain fall, they are not getting proper yield, and also expected price for their crops. The farmers are getting loss or less profit. If they know the price of particular commodity, accordingly they can grow selected crop depending upon their geographical area. We can use some of the machine leaning regression algorithms to predict the price of the agricultural commodities that helps the formers to do the decision making. This research work focus on detailed study of predicting the price of the cotton using five different ML regression algorithms namely Linear Regression, Bayesian Linear Regression, Boosted Decision Tree Regression, Decision forest Regression, Poisson Regression. Metrics for evaluating the performance of ML model. The Data set contains the cotton prices of different states of India. The cotton price 2019 data set is used for this work and the prediction values are compared with 2020 data set and the predicted prices are nearing to the actual values. Compare to all other algorithms Boosted decision tree giving the good accuracy, after that Bayesian and Linear regression also performing well.