Title: Performance analysis of data mining software with parametric changes

Authors: Ercan Atagün; Irem Duzdar Argun

Addresses: Department of Computer Engineering, Duzce University, Duzce, Turkey ' Department of Industrial Engineering, Duzce University, Duzce, Turkey

Abstract: There are significant increases in data mining software as in the data mining studies in the last years. Those increasing software are developed depending on the usage areas. Because of this reason, various parameters are supplied to the users. The parameters have the default values, but they may change with the values used by the user, and they are integrated when the algorithm is applied. In this study, supplying various parameters according to the classification parameters of the data mining software, and the effect of those parameters, which are manageable, on the success of the classification are analysed. Three different algorithms, support vector machine (SVM), K nearest neighbour (KNN) and artificial neural networks (ANN) are applied to five various data mining and the effects of manageable parameters on the classification successes are studied.

Keywords: data mining; software; classification; parameter; performance analysis.

DOI: 10.1504/IJFSE.2020.110624

International Journal of Forensic Software Engineering, 2020 Vol.1 No.2/3, pp.115 - 143

Received: 14 Feb 2020
Accepted: 25 Jun 2020
Published online: 26 Oct 2020 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article