IMPLEMENTATION OF DECISION TREE AND SUPPORT VECTOR MACHINE ON RAISIN SEED CLASSIFICATION

Wardhani Utami Dewi (Universitas Lampung)
Khoirin Nisa (Universitas Lampung)
Mustofa Usman (Universitas Lampung)

AKSIOMA: Jurnal Program Studi Pendidikan Matematika
Vol 12, No 1 (2023)

In everyday life there are many complex and global problems, especially in terms of decision making. Machine learning (ML) which is built from the concepts of computer science statistics and mathematics can automatically solve problems without guidance from ordinary users. Decision tree (DT) and support vector machine (SVM) are two supervised learning methods among several classification algorithms in ML. Both algorithms are the most popular classification techniques due to their ability to change a complex decision-making process into a simple process. In this study, the accuracy of the DT and SVM algorithms is studied on classifying raisin seeds into the Besni class and the Kecimen class based on existing features. The raisin data are divided into training and testing data, and the evaluation of the two methods is done using the testing data. The results of the evaluation are compared based on the accuracy, sensitivity, specificity, and kappa levels of the DT and SVM algorithms. The results on classifying raisin seeds data show that the SVM algorithm is superior to DT, therefor the number of positive observations is more precise in the prediction.