Developing Classification Model for Chickpea Types using Machine Learning Algorithms
Nigus Asres Ayele1, Hailemichael Kefie Tamiru2

1Nigus Asres Ayele, Information Technology, Wolkite University, Wolkite, Ethiopia.
2Hailemichael Kefie Tamiru, Software Engineering, Wolkite University, Wolkite, Ethiopia.

Manuscript received on September 20, 2020. | Revised Manuscript received on November 03, 2020. | Manuscript published on November 10, 2021. | PP: 5-11 | Volume-10 Issue-1, November 2020 | Retrieval Number: 100.1/ijitee.A80571110120| DOI: 10.35940/ijitee.A8057.1110120
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)

Abstract: Ethiopia is the leading producer of chickpea in Africa and among the top ten most important producers of chickpea in the world. Debre Zeit Agriculture Research Center is a research center in Ethiopia which is mandated for the improvement of chickpea and other crops. Genome enabled prediction technologies trying to transform the classification of chickpea types and upgrading the existing identification paradigm. Current state of the identification of chickpea types in Ethiopia still sticks to a manual. Domain experts tried to recognize every chickpea type, the way and efficiency of identifying each chickpea types mainly depend on the skills and experience of experts in the domain area and this frequently causes error and sometimes inaccurate. Most of the classification and identification of crops researches were done outside Ethiopia; for local and emerging varieties, there is a need to design classification model that assists selection mechanisms of chickpea and even accuracy of an existing algorithm should be verified and optimized. The main aim of this study is to design chickpea type classification model using machine learning algorithm that classify chickpea types. This research work has a total of 8303 records with 8 features and 80% for training and 20% for testing were used. Data preprocessing were done to prepare the dataset for experiments. ANN, SVM and DT were used to build the model. For evaluating the performance of the model confusion matrix with Accuracy, Recall and Precision were used. The experimental results show that the best-performed algorithms were decision tree and achieve 97.5% accuracy. After the evaluation of results found in this research work, agriculture research centers and companies have benefited. The model of chickpea type classification will be applied in Debre Zeit agriculture research center in Ethiopia as a base to support the experts during chickpea type identification process. In addition it enables the expertise to save time, effort and cost with the support of the identification model. Moreover, this research can also be used as a corner stone in the area and will be referred by future researchers in the domain area. 
Keywords: Chickpea, Phenotype, Varieties, Identification, classification, Selection.