A Disparate Univariate Regression Model for Sugarcane Crop Forecasting to Meet the Global Demand

Abstract

The art of predicting the yield of crop before harvesting is termed as crop forecasting. Crop forecasting helps in economic planning and also ensures global food security. Crop forecasting depends on various factors such as crop management, plant physiology, meteorology, soil science and statistical models etc. The dataset used for this study contains one independent variable hectares of land in which sugarcane is cultivated and one dependent variable yield in tones. This paper proposed a Disparate Univariate Regression Model (DURM) which uses simple linear, polynomial, Decision Tree,
Random Forest and Support Vector Regression to predict the yield of sugarcane crop with respect to the hectares of land used for sugarcane cultivation. The Regression Models were built and the Equation was framed with x and y intercepts values and the future yield of sugarcane was predicted. The correlation between the hectares of the land cultivated and sugarcane produced in tones was calculated using the correlation metrics and the performance of the various regression
models was tested with the error metrics such as (RMSE, MAE, MSE, MAPE) and the results were compared. The extracted knowledge through thisstudy helps the farmers to take decision related to their business activities and it also helps to improve food security.