Abstract | ||
---|---|---|
•A hybrid architecture composed of an LSTM and a soft GBDT is introduced.•Joint optimization for feature extraction and decision making is employed.•Negative effects of separate feature and model selection are addressed.•Components of the architecture can be generically replaced with parallel models.•Significant performance improvements in regression are achieved. |
Year | DOI | Venue |
---|---|---|
2022 | 10.1016/j.dsp.2022.103687 | Digital Signal Processing |
Keywords | DocType | Volume |
Feature extraction,End-to-end learning,Online learning,Prediction,Long short-term memory (LSTM),Soft gradient boosting decision tree (sGBDT) | Journal | 129 |
ISSN | Citations | PageRank |
1051-2004 | 0 | 0.34 |
References | Authors | |
0 | 2 |
Name | Order | Citations | PageRank |
---|---|---|---|
Mustafa E. Aydın | 1 | 0 | 0.34 |
Suleyman S. Kozat | 2 | 11 | 2.28 |