en-de
en-es
en-fr
en-sl
en
en-zh
0.25
0.5
0.75
1.25
1.5
1.75
2
Neighbourhood Components Analysis
Published on Feb 25, 200719670 Views
Say you want to do K-Nearest Neighbour classification. Besides selecting K, you also have to chose a distance function, in order to define "nearest". I'll talk about a novel method for *learning* -- f
Related categories
Chapter list
NEIGHBOURHOOD COMPONENTS ANALYSIS00:06
Basic Classifiers Perform Annoyingly Well01:00
K Nearest Neighbour Classification01:50
Problems with Nearest Neighbour02:48
Learning a Distance Metric04:29
Cross-Validation Performance is Hard to Optimize06:20
Stochastic Neighbour Selection07:44
Expected Leave-One-Out Error09:19
Quadratic Metrics10:13
Optimizing Expected Performance12:04
Neighbourhood Components Analysis (NCA)13:56
Scale of A is also learned15:16
Low Rank Metric - Nonsquare A16:29
Illustration: Concentric Rings18:41
Toy Data: UCI Wine19:53
Face Data21:05
Summary of NCA23:05
Scale of A is also learned 0128:18
Face Data 0130:34