 |
International Society of Science and Applied Technologies |
| Overview of One-Class versus Binary-Class Classification in the Context of Big Data Analytics | ||||
| Author | Naeem Seliya
|
|||
| Co-Author(s) | Taghi M. Khoshgoftaar
|
|||
| Abstract | The primary focus of this article is to provide an insight into the state of the literature where a direct comparison between one-class classification and binary class classification is conducted. The latter is by-far the most common type of categorical problem addressed by experts in the machine learning community. In one-class classification, a supervised model is built based on instances from only one class, and then prediction is done in the form of inliers and outliers. Throughout this article, we highlight the differences between these two types of classification approaches, especially in the case of the severe class imbalance problem and the class rarity problem. Moreover, we investigate the literature and discuss whether one-class classification has been addressed with big data for the problems of severe class imbalance and class rarity. We conclude, to the best of our knowledge, that there are no studies that have done so.
|
|||
| Keywords | One-class classification, binary-class classification, class imbalance, class rarity, big data | |||
| Article #: RQD28-391 | ||||
Proceedings of 28th ISSAT International Conference on Reliability & Quality in Design |