Regularization of Kernel Methods by Decreasing the Bandwidth of the Gaussian Kernel thumbnail
slide-image
Pause
Mute
Subtitles not available
Playback speed
0.25
0.5
0.75
1
1.25
1.5
1.75
2
Full screen

Regularization of Kernel Methods by Decreasing the Bandwidth of the Gaussian Kernel

Published on Jun 11, 20076187 Views

Related categories

Chapter list

Regularization of Kernel Methods by Decreasing the <br>Bandwidth of the Gaussian Kernel00:00
Outline01:12
Gaussian kernel and RKHS01:25
Gaussian kernel and RKHS 0102:47
Gaussian RKHS03:11
Learning in Gaussian RKHS04:47
Motivation 1: The effect of regularization06:04
Motivation 1: The effect of regularization 0114:37
Motivation 2: One-class SVM16:17
Outline21:37
Setting and notations21:57
Intuitive behavior23:44
Intuitive behavior 0125:15
Main result: consistency25:55
Main result: consistency 0127:02
Main result: asymptotic shape27:36
Main result: asymptotic shape 0128:01
Application: two-class SVM28:27
Application: two-class SVM 0129:17
Application: one-class SVM29:18
Conclusion30:41