Classification for Optimization Accuracy Dry Bean (Phaseolus vulgaris L.) using Random Forest Algorithm with Python
Rohmat Taufiq (a*), Adi Wibowo (b), Budi warsito (c)

a) Doctoral Program of Information System, School of Postgraduate Studies, Diponegoro University, Semarang, Indonesia.

a) Informatic Engineering Department, University Muhammadiyah Tangerang, jl. Perintis Kemerdekaan I/33 Cikokol, Tangerang, Indonesia

b) Departement of Computer Science, Informatics, Diponegoro University, Semarang, Indonesia.

c) Departement of Statistics, Faculty Sciences and Mathematics, Diponegoro University, Semarang Indonesia.

Abstract

Agriculture is a sector that includes food crop cultivation, animal husbandry, dairy farming, fisheries, forestry, and other related activities that promise to grow the economy of the poor. Red beans have a very complex nutritional content, which has many benefits for the body, especially for people with diabetes, stroke, high cholesterol, and similar diseases. There is the various genetic diversity of dried beans in the world, all of which is one factor influenced by the quality of the species. Therefore, seed classification is very important both for marketing and production to be able to make a sustainable farming system. This research uses Python and Dry Bean Dataset obtained from UCI will be compared between the Random Forest Algotirma with Gradient Boosting, Extra Trees, Bagging, XGB, and LGBM Classifier. This research concludes that the accuracy from the highest to the lowest values starts from the LGBM classifier 92.65%, XGB 92.35%, Random Forest 92.29%, Gradient Boosting 91.91%, Extra Trees 91.67%, Bagging 91.67%. Extra Trees and Bagging have the same value of 91.67%. the criteria that have the highest influence are Shape Factor 4, followed by Roundness, Solidity to the last Eccentricity. As a suggestion for further research, it is necessary to compare it with other methods and combine it with other optimization methods to get a higher accuracy value.

Keywords: classification- optimization- dry bean- lgbm- xgb- random forest- gradient boosting- extra trees- bagging

Topic: Computer Science