Comparison of Classification Results of Smo and J48 Algorithms on Different Data Sets

Abstract

Aim: Data mining is an interdisciplinary field, constantly developing and expanding its use. It helps to ensure the reliability of data by using various techniques and algorithms. Classification is an important data mining technique because it is widely used by researchers. Method: In this study, the classification results of SMO and J48 algorithms were compared on the data of three arc students. The performance of J48 and SMO algorithms in terms of classification accuracy was evaluated using three different data sets and various accuracy measurements such as TP-Rate, FP-Ratio, Precision, Precision, F-criterion and ROC analysis. Findings and Results: As a result of the tests, it was revealed that the classification performance of SMO algorithm was better in all three datasets.

Determining the Expressions on the Employee Performance Scale With the Decision Tree Algorithm

“…Bu kriterler bilgi kazancı ve bilgi kazancı oranı, giniindeksi, Towing kuralı, Ki -Kare olasılık tablo istatistiği olarak sayılabilmektedir. Bu yöntemler kullanılarak geliştirilen algoritmalardan bazıları ise ID3, C4.5, C5.0, J48 (Alan ve Yeşilyurt, 2018) ve CART algoritmaları olarak sayılabilmektedir (Kavzoğlu ve Çölkesen, 2010). Bu çalışmada bilgi kazancı oranı yaklaşımı kullanılarak tasarlanan J48 algoritması kullanılmıştır.…”