Cross-domain topic transfer learning method based on multiple balance and feature fusion
In transfer learning, traditional homogeneous transfer learning assumes similar data and feature distributions between the source and target domains, focusing primarily on parameter sharing to enhance model performance. However, heterogeneous transfer learning for topic model, disparities in data an...
Saved in:
| Main Authors: | , , , |
|---|---|
| Format: | Article |
| Language: | English |
| Published: |
Elsevier
2025-05-01
|
| Series: | Heliyon |
| Online Access: | http://www.sciencedirect.com/science/article/pii/S2405844024167941 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|