Jump to content

Talk:Von Mises–Fisher distribution

Page contents not supported in other languages.
From Wikipedia, the free encyclopedia

Presumably it generalizes to a Mises distribution (circle) for p=3, not p=2, since circles are 2-dimensional and the von Mises-Fisher distribution is p-1 dimensional. There might be something I missed though, so I'm not going to change it without confirmation. --Smári McCarthy 13:16, 19 July 2006 (UTC)[reply]

The circle is not 2-dimensional, but 1-dimensional, as any point on the circle can be uniquely identified by a singe angle. Similarly, the sphere is 2-dimensional, as each point can be identified by a pair of polar angles. However, you do need an extra dimension to 'represent' the circle or sphere without distortions. --TomixDf Mon Aug 7 12:02:13 2006

The R notation is undefined

[edit]

what is it ? — Preceding unsigned comment added by 92.133.97.155 (talkcontribs)

I think the "mean resultant vector". See Directional statistics. --Tobias1984 (talk) 14:13, 29 July 2015 (UTC)[reply]

Someone please clarify which von Mises

[edit]

Was it Ludwig von Mises, the Austrian economist? The article doesn't say, nor does the article on the von Mises distribution. I was surprised recently to find out that LvM had in fact done early work on algorithmic randomness. --Trovatore (talk) 01:17, 23 February 2008 (UTC)[reply]

It was his brother, Richard von Mises. Tomixdf (talk) 09:04, 23 February 2008 (UTC)[reply]

Mardia reference missing

[edit]

What is the Mardia (2000) reference? — Preceding unsigned comment added by 128.243.253.117 (talk) 13:03, 13 June 2011 (UTC)[reply]

Polar coordinates?

[edit]

The article first states that x is a p-dimensional unit vector. Then there is a comment that says "Note that the equations above apply for polar coordinates only.". I don't believe this is the case? The references seem to refer to x in R^(p). If I've missed something, then surely at the least these are hyperspherical coordinates, not polar?

I tried testing this in a very simple 2D (circle) case, and using cartesian coordinates gave me the expected result --TheKrimsonChin (talk) 15:35, 27 February 2015 (UTC)[reply]

Indeed the reported functions are for cartesian, not polar(spherical in reality) coordinates, I'm going to fix it. [Silvano Galliani] — Preceding unsigned comment added by 192.33.89.33 (talk) 08:11, 26 June 2015 (UTC)[reply]