Curse of Dimensionality

preview_player
Показать описание
This is a video about a nasty topic in data-analysis: The curse of dimensionality. It is quite hard to get your head around it because it is inherently a high-dimensional phenomenon that in many respects doesn't have a low dimensional analog.

The name of the curse was originally coined by Richard Bellmann in his book on Dynamic Programming (Princeton University Press 1957).

In time the curse has taken on several meanings of which three are discussed here. These refer to the "volume", "distances" and "hubness" of the data.

More information on the last point of Hubness which I only mention briefly can be found in: Radovanovic et al. Journal of Machine Learning 11, 2487-2531 (2010).

CONTENTS
0:00 Introduction
0:53 Volume of data
5:15 Distance distribution
8:33 Hubness
9:05 Final words
Рекомендации по теме
Комментарии
Автор

That's a really interesting subject. I greatly appreciate you sharing such valuable content with anyone who is interested in this area.

azraseyyedi