video thumbnail
Pause
Mute
Subtitles
Playback speed
0.25
0.5
0.75
1
1.25
1.5
1.75
2
Full screen

Exploring the hubness-related properties of oceanographic sensor data

Published on 2011-11-042775 Views

In this paper we examine how the high dimensionality of oceanographic sensor data impacts the potential use of nearest-neighbor machine learning methods. We focus on one particular consequence of the

Related categories

Presentation

Exploring the hubness-related properties of oceanographic sensor data00:00
Presentation outline00:25
Why it matters - 100:43
Why it matters - 201:17
Related work02:05
Analysis goals02:35
Test case: Oceanographic data - 102:38
Test case: Oceanographic data - 203:15
The experimental setup04:07
Skewness, bad hubness05:11
Class to class hubness matrix, k=3, wind measurements06:05
Would the hubness-aware methods help?07:08
Wind measurements: Sensor hubness map - 107:29
Wind measurements: Sensor hubness map - 207:52
Water temperature: Sensor hubness map - 108:35
Water temperature: Sensor hubness map - 208:39
Barometric pressure: Sensor hubness map08:52
Air temperature: The bermuda triangle08:56
Conclusions09:21
Thank you for your attention09:24