Forecasting the Factors Responsible for Improving the Yield of Sugarcane Crop using Artificial Neural Network

Abstract

Sugarcane is a cash crop cultivated in India's Tropical and Sub Tropical regions, contributing 5.7% of the Gross Domestic Product (GDP) to the Indian Economy. Sugarcane Farming gives employment opportunities to 60 million rural families. It is cultivated all over India to a latitude of 80◦ N to 300◦ N. Sugarcane contains sucrose in its stem juice and is the primary raw material for producing sugar. Sugarcane crop production depends on the season, biological, and economic cause. The growing demand for sugarcane worldwide incorporates the backbone of sugarcane agriculture. This paper proposed A Hybrid Machine Learning Model (HMLM) for forecasting the sugarcane yield, which helps both farmers and the sugar mills to assist in annual planning. The proposed model used Backward Feature Elimination to select the factors that dominate sugarcane production. K-Means Clustering (K-MC) is applied to the selected attributes, and the dataset is partitioned. An Artificial Neural Network (ANN) is created for each clustered dataset, and the parameters influencing sugarcane production are found. The performance of each Neural Network (NN) created was analysed by implementing performance measures, and the findings were compiled. The information gained from the suggested model aids farmers in making decisions to increase sugarcane crop productivity.