Name
Abstract | ||
|---|---|---|
This paper investigates a new learning setting recently introduced by Vapnik [8] that takes into account a known structure of the training data to improve generalization performance. This setting is a special case of a new inference technology known as Learning with Hidden Information[8] suitable for many real-life applications with sparse high-dimensional data. We first briefly describe an extension of SVM called SVM gamma+ [8] that is associated with this new learning setting, and verify its effectiveness using a synthetic data set. Then we demonstrate the effectiveness of SVM gamma+ on a difficult real-life problem: detection of cognitive states from AM images obtained from different subjects. These empirical results show that the SVM gamma+ approach achieves improved inter-subject generalization vs standard SVM technology. |
Year | DOI | Venue |
|---|---|---|
2007 | 10.1109/IJCNN.2007.4371006 | 2007 IEEE INTERNATIONAL JOINT CONFERENCE ON NEURAL NETWORKS, VOLS 1-6 |
Keywords | Field | DocType |
synthetic data,data analysis,support vector machines,structured data,high dimensional data,learning artificial intelligence | Online machine learning,Semi-supervised learning,Instance-based learning,Stability (learning theory),Active learning (machine learning),Pattern recognition,Computer science,Support vector machine,Unsupervised learning,Artificial intelligence,Computational learning theory,Machine learning | Conference |
ISSN | Citations | PageRank |
2161-4393 | 11 | 0.91 |
References | Authors | |
1 | 2 | |
Name | Order | Citations | PageRank |
|---|---|---|---|
| Lichen Liang | 1 | 80 | 7.39 |
| Vladimir Cherkassky | 2 | 1064 | 126.66 |