Bánhalmi Árpád (2023) Hierarchical Clustering Combined with Neural Networks. Business & Diplomacy Review, 1 (2). pp. 7-21. ISSN 3004-0116
Preview |
Text
BDR-23-2-Bánhalmi.pdf - Published Version Download (2MB) | Preview |
Abstract
This study presents how hierarchical clustering combined with neural networks operates and sheds light on how these models can be integrated with fundamental concepts. The hybrid model created can be applied to various tasks depending on the configuration of hyperparameters. The study serves a demonstration purpose by showing how the suggested model performs when run on a real dataset, where it attempts to estimate ordinal target variable values. The model’s performance is analyzed in detail, with particular emphasis on overfitting, stability, and estimation accuracy. The hybrid model generates multiple estimates for the target variable values simultaneously - in this case, ten. These estimates are compared using available metrics, and then a method for deriving the final estimate is presented.
Tudományterület / tudományág
természettudományok > matematika- és számítástudományok
Kar
Institution
Budapesti Gazdasági Egyetem
| Item Type: | Article | ||||||
|---|---|---|---|---|---|---|---|
| Creators: |
|
||||||
| Uncontrolled Keywords: | neural network, hierarchical clustering, overfitting management | ||||||
| Depositing User: | Erzsébet Kovály | ||||||
| Identification Number: | 34505334 | ||||||
| Date Deposited: | 2024. Feb. 22. 12:12 | ||||||
| Last Modified: | 2025. May. 16. 08:58 | ||||||
| URI: | https://publikaciotar.uni-bge.hu/id/eprint/2271 |
Actions (login required)
 |
View Item |
