پیوندهای مفید
اخبار و اعلانات
آمار
| تعداد نشریات | 127 |
| تعداد شمارهها | 7,240 |
| تعداد مقالات | 77,673 |
| تعداد مشاهده مقاله | 161,286,311 |
| تعداد دریافت فایل اصل مقاله | 120,868,229 |
A Critical Examination of Distributed Relief-Based Ensemble Clustering Frameworks for Time-Series Analysis | ||
| Journal of Algorithms and Computation | ||
| دوره 58، شماره 1، مهر 2026، صفحه 131-145 اصل مقاله (1.72 M) | ||
| نوع مقاله: Research Paper | ||
| شناسه دیجیتال (DOI): 10.22059/jac.2026.414734.1262 | ||
| نویسنده | ||
| Ainaz Bahramlou* | ||
| Department of Electrical and Computer Engineering, Isfahan University of Technology, Isfahan 84156-83111, Iran | ||
| چکیده | ||
| Ensemble clustering methods have garnered significant attention for their ability to improve robustness and accuracy in time-series analysis. This paper critically examines two recently proposed parallel and distributed ensemble clustering frameworks leveraging MapReduce and Relief-family feature weighting (including ReliefF, Multisurf*, Simba-Sc, I-Relief, and M-Relief). While these frameworks offer advancements in computational efficiency for large datasets, our analysis reveals significant limitations in their methodological positioning and evaluation validity. We contend that these approaches do not constitute deep learning models, as they fundamentally lack hierarchical representation learning, end-to-end differentiable optimization, and the learning of latent representations. Furthermore, the reported performance metrics, such as runtime, memory usage, clustering accuracy, diversity, and generic error measures, primarily reflect algorithmic efficiency under controlled conditions rather than genuine predictive reliability in complex, real-world scenarios. We highlight the absence of robust uncertainty quantification, specifically aleatoric and epistemic uncertainty, which is crucial for reliable decision-making. | ||
| کلیدواژهها | ||
| Ensemble Clustering؛ Time-Series Analysis؛ Relief Feature Weighting؛ Distributed Computing؛ MapReduce؛ Deep Learning؛ Uncertainty Quantification؛ Fuzzy Logic؛ Neural Data Analysis | ||
|
آمار تعداد مشاهده مقاله: 47 تعداد دریافت فایل اصل مقاله: 29 |
||
سامانه مدیریت نشریات علمی. قدرت گرفته از سیناوب