Districtwise Economic Analysis of Sugarcane Farming in Madhya Pradesh using Machine Learning: A Comprehensive Assessment

Abstract

This study focuses on predicting sugarcane farming areas in different districts of Madhya Pradesh using two distinct models, Linear Regression and ARIMA. The primary objective is to compare the performance of the Linear Regression and ARIMA models in forecasting sugarcane farming areas. The analysis begins by preprocessing the dataset, removing irrelevant data, and splitting it into training and testing sets. The Linear Regression model is employed to learn the linear relationship between input features, such as district-wise productivity data, and the target variable, sugarcane farming area. Subsequently, the model predicts productivity values based on the training data. Additionally, the ARIMA model, a time series forecasting method, is implemented to capture the temporal patterns in the sugarcane farming data. It takes into account the seasonal and trend components in the time series to produce predictions. The evaluation of the models is performed based on mean squared error (MSE) and mean absolute error (MAE) metrics. The findings reveal that the Linear Regression model performs better than the ARIMA model in this specific prediction task. It yields predictions that are more accurate and closer to the actual sugarcane farming area values. Overall, the study demonstrates the effectiveness of Linear Regression as a predictive tool for estimating sugarcane farming areas in Madhya Pradesh. The results can provide valuable insights for agricultural planning and resource allocation in the region, potentially aiding policymakers and farmers to make informed decisions and enhance agricultural productivity in the future.