Data Mining Tools Evaluation Based on their Quality Attributes

  • Rami Malkawi, Ahmad A. Saifan, Nouh Alhendawi, Alaa Bani Ismaeel

Abstract

As a result of the rapid evolution of open source software, the entire software industry has been grown up in a way that made it harder to improve the software quality or even keep it as it is without any shortages in the quality level. However, data mining tools are considered as a good example of open source software, which has a big similarity between them such as in goals, results and maybe in the data set which are used in work, but for sure they are different in terms of quality. Quality attributes such as maintainability, reusability and fault-proneness are important factors in any process. In addition, they can tell which tool is appropriate for specific jobs. In this paper the quality of five open source data mining products (Weka 3.9, Rapid Miner, Knime, Apache Mahout and Keel) is investigated by studying their quality metrics and attribute, to help the software developers and researchers decide and select the best data mining tools that meet their desired needs.  By applying a predefined maintainability index, ApacheMahout tool has been found the most maintainable. Also, Keel had the most reusable components based on the reusability index formula used in this paper.  As a result of this research, a formula (according to a previous equation) has been developed, based on object oriented metrics, to measure the fault-proneness of the selected data mining tools.

Published
2020-03-30
How to Cite
Rami Malkawi, Ahmad A. Saifan, Nouh Alhendawi, Alaa Bani Ismaeel. (2020). Data Mining Tools Evaluation Based on their Quality Attributes. International Journal of Advanced Science and Technology, 29(3), 13867 - 13890. Retrieved from http://sersc.org/journals/index.php/IJAST/article/view/31731
Section
Articles