Dry Beans Classification using Ensemble Learning

Abstract

One of the highest-produced crops in the world, Dry Bean faces an extreme genetic diverse species in their crops. The quality of the seed is influencing in production of crops. Consequently, the classification of the seeds has become the need of the hour for production as well as marketing in order to avail the agricultural principles of sustainable systems. This research aims to develop a method which helps to obtain even varieties of seeds from the production of crops. In the literature, very few works have been developed for the Dry beans classification. In this work, an ensemble model for the classification called ELDB is developed where ELDB stands for Ensemble Learning classifier for Dry Beans. The proposed method uses the philosophy of ensemble of ensembles to develop a robust classifier to classify Dry beans effectively. Based on the different extracted features, ELDB is trained, an ensemble of Random Forest and XGboost. The proposed ensemble model has the highest performance score compared to other methods with class-wise accuracies such as Seker, Barbunya, Bombay, Cali, Horoz, Sira and Dermason with 96.94%, 96.06%, 91.95%, 96.32%, 96.16%, 95.9% and 98.42% respectively. The overall performance of the proposed method has been compared with the state-of-art method over various measures like accuracy, precision, recall and F1-score, where the proposed method performance are 95.96%, 95.71%, 95.84% and 95.97% respectively.