Adaptive Feature Selection of Unbalanced Data for Skiing Teaching
In skiing teaching, the data may show an unbalanced distribution. For example, the sample size of some common movements (such as straight downhill) may be much larger than that of some difficult movements (such as aerial spins). If the features are not selected, the model may overly rely on the feat...
Saved in:
| Main Author: | |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Tamkang University Press
2025-06-01
|
| Series: | Journal of Applied Science and Engineering |
| Subjects: | |
| Online Access: | http://jase.tku.edu.tw/articles/jase-202603-29-03-0006 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | In skiing teaching, the data may show an unbalanced distribution. For example, the sample size of some common movements (such as straight downhill) may be much larger than that of some difficult movements (such as aerial spins). If the features are not selected, the model may overly rely on the features of common actions and ignore the features of difficult actions. Through the feature selection of unbalanced data, features that have significant influences on different actions (especially a few types of actions) can be screened out, thereby enhancing the recognition ability and generalization ability for various actions. The high-dimensional characteristics of the skiing dataset will reduce the classification effect of unbalanced learning. Aiming at the
classification problem of high-dimensional unbalanced data, this paper proposes an adaptive feature selection method. This new algorithm combines embedded and wrapped feature selection methods and is capable of adaptively selecting the optimal features to form the feature space. Finally, the experimental results on the public imbalanced dataset show that the proposed algorithm effectively improves the classification performance of imbalanced data. |
|---|---|
| ISSN: | 2708-9967 2708-9975 |