A Rule Based Recommender System to Improve the Yield of Groundnut Crop Using Decision Tree with Backward Elimination, Principal Component Analysis

Abstract

Precision farming is a new technological revolution that is presently happening in the agriculture field
which aims at making an individual farmer to produce food for 155 people. Practicing precision farming push farmers to adopt new cropping technology that uses advanced software and IOT sensor equipped devices. In precision farming data is generated from various sensors in huge volume which requires special storage mechanism and machine learning algorithms to analyse the data. Decision Tree built using high dimension datasets takes more time to construct the tree, requires huge memory space and also produce complex rules. This paper proposed a machine learning model that has two decision tree algorithms namely Decision Tree with Principal Component Analysis (DTPCA) and Decision Tree with Backward Elimination (DTBE) that combines the Decision Tree Algorithm with feature selection techniques such as Principal Component Analysis (PCA), and Backward Elimination. This proposed framework aims to develop a rule based recommender system that will assist farmers in decision making and helps to improve the yield of Groundnut. The proposed algorithm was tested with a real time dataset that contains factors responsible for the growth of groundnut yield. The results
showed that the Decision Tree Algorithm combined with Principal Component Analysis performs better and classify the dataset with higher accuracy and low error rate