Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Both sides previous revision Previous revision
default:oscpls [2017/03/07 08:58]
admin
default:oscpls [2017/03/07 08:59] (current)
admin
Line 1: Line 1:
 +====== OSC-PLS-DA ======
 +
 +== Description ==
 +
 +Partial least squares regression (PLS regression) is a statistical method that bears some relation to principal components regression; instead of finding hyperplanes of minimum variance between the response and independent variables, it finds a linear regression model by projecting the predicted variables and the observable variables to a new space. Because both the X and Y data are projected to new spaces, the PLS family of methods are known as bilinear factor models. Partial least squares Discriminant Analysis (PLS-DA) is a variant used when the Y is categorical.See [[http://​en.wikipedia.org/​wiki/​Partial_least_squares_regression|Wikipedia]].
 +
 +But, because PLS has problems in dealing with strong and structured noise in the descriptor matrix X, the Orthogonal Signal Correction (OSC) filters can be used to remove structured Y-orthogonal variation from X. Cf [[main:​references|Wehrens R. (2011)]]
 +
 +__See also__:
 +   * [[https://​imdevsoftware.wordpress.com/​2013/​03/​15/​evaluation-of-orthogonal-signal-correction-for-pls-modeling-osc-pls-and-opls/​|evaluation-of-orthogonal-signal-correction-for-pls-modeling-osc-pls-and-opl]]
 +
 +\\ 
 +----
 +\\ 
 +
 +== Usage ==
 +
 +^Field ^Description ^
 +|Launch | Field indicating whether you want to execute this analysis in the workflow (checked) or not (unchecked).|
 +//to finish ...// 
 +\\ 
 +----
 +\\ 
 +
 +== Output Examples ==
 +
 +<columns 100% 50% left top ->
 +
 +**F1 OSC-PLS**
 +{{ ::​f1_osc-pls.png?​direct&​300 |}} 
 +
 +<​newcolumn>​
 +
 +**F1 Validation**
 +{{ ::​f1_validation.png?​direct&​300 |}}
 +
 +<​code>​
 +Data: X dimension: 54 191 
 + Y dimension: 54 1
 +Fit method: kernelpls
 +Number of components considered: 7
 +
 +----------------------------------
 +
 +[1] "R2 Y,X = 0.22965262"​
 +
 +                 ​PLS ​      ​OSC1 ​      OSC2
 +R2 Yhat,X 0.23037326 0.23021366 0.23003459
 +R2 Y,Xm   ​0.28590360 0.38412967 0.43262628
 +R2 to,X           NA 0.20987811 0.10139296
 +R2 t1,X   ​0.25413110 0.24079471 0.23426993
 +R2 t2,X   ​0.19654173 0.09486818 0.03244769
 +R2 t1,Xm  0.31741439 0.39273250 0.44036621
 +R2 t2,Xm  0.24635207 0.16870403 0.06984592
 +R2 Y,t1   ​0.92481714 0.96073283 0.98150053
 +R2 Y,t2   ​0.03591568 0.02076770 0.01087043
 +</​code>​
 +</​columns>​
 +
 +<columns 100% 50% left top  ->
 +
 +**F2 OSC-PLS**
 +{{ ::​f2_osc-pls.png?​direct&​300 |}} 
 +
 +<​newcolumn>​
 +
 +**F2 Validation**
 +{{ ::​f2_validation.png?​direct&​300 |}}
 +
 +<​code>​
 +Data: X dimension: 54 191 
 + Y dimension: 54 1
 +Fit method: kernelpls
 +Number of components considered: 13
 +
 +----------------------------------
 +
 +[1] "R2 Y,X = 0.18532673"​
 +
 +                PLS       ​OSC1 ​      OSC2
 +R2 Yhat,X 0.1854653 0.18540961 0.18538841
 +R2 Y,Xm   ​0.2081634 0.27796614 0.33316421
 +R2 to,X          NA 0.19488805 0.17253656
 +R2 t1,X   ​0.2419817 0.20145188 0.19307770
 +R2 t2,X   ​0.1543582 0.16416238 0.08040183
 +R2 t1,Xm  0.2712700 0.30044413 0.34878165
 +R2 t2,Xm  0.1710830 0.21808801 0.15893459
 +R2 Y,t1   ​0.8237835 0.93760281 0.96324369
 +R2 Y,t2   ​0.1138193 0.02564088 0.01264715
 +</​code>​
 +</​columns>​
 +
 +
 +
  
default/oscpls.txt ยท Last modified: 2017/03/07 08:59 by admin