Details
| Originalsprache | Englisch |
|---|---|
| Titel des Sammelwerks | Advances in Visual Computing |
| Untertitel | 10th International Symposium, ISVC 2014, Proceedings |
| Herausgeber (Verlag) | Springer Verlag |
| Seiten | 95-106 |
| Seitenumfang | 12 |
| ISBN (elektronisch) | 9783319143637 |
| Publikationsstatus | Veröffentlicht - 2014 |
| Veranstaltung | 10th International Symposium on Visual Computing, ISVC 2014 - Las Vegas, USA / Vereinigte Staaten Dauer: 8 Dez. 2014 → 10 Dez. 2014 |
Publikationsreihe
| Name | Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) |
|---|---|
| Band | 8888 |
| ISSN (Print) | 0302-9743 |
| ISSN (elektronisch) | 1611-3349 |
Abstract
The original Random Forest derives the final result with respect to the number of leaf nodes voted for the corresponding class. Each leaf node is treated equally and the class with the most number of votes wins. Certain leaf nodes in the topology have better classification accuracies and others often lead to a wrong decision. Also the performance of the forest for different classes differs due to uneven class proportions. In this work, a novel voting mechanism is introduced: each leaf node has an individual weight. The final decision is not determined by majority voting but rather by a linear combination of individual weights leading to a better and more robust decision. This method is inspired by the construction of a strong classifier using a linear combination of small rules of thumb (AdaBoost). Small fluctuations which are caused by the use of binary decision trees are better balanced. Experimental results on several datasets for object recognition and action recognition demonstrate that our method successfully improves the classification accuracy of the original Random Forest algorithm.
ASJC Scopus Sachgebiete
- Mathematik (insg.)
- Theoretische Informatik
- Informatik (insg.)
Zitieren
- Standard
- Harvard
- Apa
- Vancouver
- BibTex
- RIS
Advances in Visual Computing: 10th International Symposium, ISVC 2014, Proceedings. Springer Verlag, 2014. S. 95-106 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Band 8888).
Publikation: Beitrag in Buch/Bericht/Sammelwerk/Konferenzband › Aufsatz in Konferenzband › Forschung › Peer-Review
}
TY - GEN
T1 - Thresholding a Random Forest classifier
AU - Baumann, Florian
AU - Li, Fangda
AU - Ehlers, Arne
AU - Rosenhahn, Bodo
N1 - Funding information: This work has been partially funded by the ERC within the starting grant Dynamic MinVIP.
PY - 2014
Y1 - 2014
N2 - The original Random Forest derives the final result with respect to the number of leaf nodes voted for the corresponding class. Each leaf node is treated equally and the class with the most number of votes wins. Certain leaf nodes in the topology have better classification accuracies and others often lead to a wrong decision. Also the performance of the forest for different classes differs due to uneven class proportions. In this work, a novel voting mechanism is introduced: each leaf node has an individual weight. The final decision is not determined by majority voting but rather by a linear combination of individual weights leading to a better and more robust decision. This method is inspired by the construction of a strong classifier using a linear combination of small rules of thumb (AdaBoost). Small fluctuations which are caused by the use of binary decision trees are better balanced. Experimental results on several datasets for object recognition and action recognition demonstrate that our method successfully improves the classification accuracy of the original Random Forest algorithm.
AB - The original Random Forest derives the final result with respect to the number of leaf nodes voted for the corresponding class. Each leaf node is treated equally and the class with the most number of votes wins. Certain leaf nodes in the topology have better classification accuracies and others often lead to a wrong decision. Also the performance of the forest for different classes differs due to uneven class proportions. In this work, a novel voting mechanism is introduced: each leaf node has an individual weight. The final decision is not determined by majority voting but rather by a linear combination of individual weights leading to a better and more robust decision. This method is inspired by the construction of a strong classifier using a linear combination of small rules of thumb (AdaBoost). Small fluctuations which are caused by the use of binary decision trees are better balanced. Experimental results on several datasets for object recognition and action recognition demonstrate that our method successfully improves the classification accuracy of the original Random Forest algorithm.
UR - http://www.scopus.com/inward/record.url?scp=84916597623&partnerID=8YFLogxK
U2 - 10.1007/978-3-319-14364-4_10
DO - 10.1007/978-3-319-14364-4_10
M3 - Conference contribution
AN - SCOPUS:84916597623
T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
SP - 95
EP - 106
BT - Advances in Visual Computing
PB - Springer Verlag
T2 - 10th International Symposium on Visual Computing, ISVC 2014
Y2 - 8 December 2014 through 10 December 2014
ER -