A New Physics-Inspired Discriminative Classifier | ||
| AUT Journal of Electrical Engineering | ||
| دوره 56، شماره 3، 2024، صفحه 495-502 اصل مقاله (625.52 K) | ||
| نوع مقاله: Research Article | ||
| شناسه دیجیتال (DOI): 10.22060/eej.2024.22694.5557 | ||
| نویسندگان | ||
| Mostafa Monemizadeh* 1؛ Seyed Rouhollah Samareh Hashemi2؛ Mohsen Sheikh-Hosseini3؛ Hamed Fehri4 | ||
| 1Department of Electrical and Computer Engineering, University of Neyshabur, Neyshabur, Iran | ||
| 2Fiber Optics Group, Institute of Science and High Technology and Environmental Sciences, Kerman, Iran | ||
| 3Department of Computer and Information Technology, Institute of Science and High Technology and Environmental Sciences, Kerman, Iran | ||
| 4Faculty of Engineering, University of Zabol, Zabol, Iran | ||
| چکیده | ||
| Concepts and laws of physics have been a valuable source of inspiration for engineers to overcome human challenges and problems. Classification is an important example of such problems that play a major role in various fields of engineering sciences. It is shown that discriminative classifiers tend to outperform their generative counterparts, especially in the presence of sufficient labeled training data. In this paper, we present a new physics-inspired discriminative classification method using minimum potential lines. To do this, we first consider two groups of fixed point charges (as two classes of data) and a movable classifier line between them. Then, we find a stable position for the classifier line by minimizing the total potential integral on the classifier line due to the two groups of point charges. Surprisingly, it will be shown that the obtained classifier is actually an uncertainty-based classifier that minimizes the total uncertainty of the classifier line. Experimental results show the effectiveness of the proposed approach. | ||
| کلیدواژهها | ||
| Classification؛ discriminative؛ machine learning؛ potential | ||
| مراجع | ||
|
| ||
|
آمار تعداد مشاهده مقاله: 619 تعداد دریافت فایل اصل مقاله: 421 |
||
| تعداد نشریات | 9 |
| تعداد شمارهها | 455 |
| تعداد مقالات | 5,771 |
| تعداد مشاهده مقاله | 8,379,045 |
| تعداد دریافت فایل اصل مقاله | 6,937,932 |