Re: PCA


sangdonlee@xxxxxxxxxxx wrote:
What Greg meant is that the first PC accounts the largest amount of
variance in the data, which is X-axis in the graph below, for example
(use the fixed courier font to see). However for classification
purpose, the second PC (Y-axis) is the better discriminator.

Greg wrote:
Don't forget that, in general, dominant PCA variable subset
selection may be inappropriate for classification.

Sangdon Lee

|
#########|#########
|
<-----------+---------->
|
#########|#########
|
--------------------------SNIP

My example was 3-D. Therefore your diagram is, with scaling,
valid for projections into both the x-z and y-z planes. Dominant
PCA chooses x and y and rejects z when what you are looking
for is just the opposite.

Or, are you just presenting a simpler 2-D example?

Hope this helps.

Greg

.

Relevant Pages

  • Re: Connecting graph vertices with edges that never intersect
    ... In 1974 Hopcroft and Tarjan published an algorithm for testing planarity ... includes Kuratowski's Theorem that gives a classification of planar ... If the graph is not planar, I think that finding a way to draw the graph ...
    (sci.math)
  • Re: KT boundry event
    ... John Harshman wrote: ... [Snip a lot of stuff irrelevant to my question] ... What other purpose does it serve, then, besides identification and ... While I disagree with the general thrust of UC's argument, I'll have to agree that any system of classification will, by its very nature, be usable as a system of identification, and vice versa. ...
    (talk.origins)
  • Classification or Pattern recognition?
    ... Are there any difference between Classification and Pattern recognition? ... I understand they are used for the same purpose but would say that Pattern ... Any other interpretations? ...
    (sci.image.processing)
Loading