Publication

Abstract : Agricultural productivity is major concern for the global food community. According to the 2011 census, about 49% of the people in India are engaged in farming or related activities. Agriculture provides almost 17% of the country's GDP. So, it's necessary to grow suitable crops at suitable locations to maximize the production of crops and revenue obtained from them to help the country’s economy. Many statistical methods have been used from past years like doing field survey and analyze the obtained data, obtaining mathematical equations that give relationship between crop-related parameters like temperature and rainfall, techniques like correlation analysis, regression analysis, ANOVA, etc. In this comparative analysis, the prediction of suitable crop is done using supervised machine learning prediction approaches such as K-Nearest Neighbor (KNN), Support Vector Machine (SVM), Least Absolute Shrinkage & Selection Operator (LASSO) regression, Linear Descriptive analysis (LDA), Nearest Centroid Classifier and Voting regression. KNN & SVM, both models achieved the highest accuracy among others, which is 97.78%. The main goal of this comparative study is that comparing different machine learning classification techniques for suitable crop prediction helps to enhance the production of crops and thus contributes to the growth of agriculture in the country. This comparative study provides valuable insights and guidance for suitable crop plantation.