Download the new release 20.90

Release 20.90
  Two new auxiliary R-functions:   Hotelling and HotellingEllipse

Description:

 Allows to detect Tē ouliers using the PCA or the PLS components and display Tē Hotelling ellipses on the same (ti, tj) scatterplot.
 Hotelling:
to detect outlying observations on the component variables ii and jj (default to 1 and 2) at the Fisher-Snedecor level (FSthreshold defaults to 0.95).
                       Hotelling displays the corresponding Tē ellipse only if FStrhesold=0.95 thus giving the initial plot.
 HotellingEllipse: When FSthreshold is different from 0.95 one have to use Hotelling and then HotellingEllipse to see the ellipse and the coloured outliers.

Example ;   PCA on 'cornell' data  (see Tenenhaus's book on pages 86-87   or  the lecture notes, page 199)
.

                                    cornellPCA=acpxqd(cornell,bgpar="white")
                                    
Hotelling(cornellPCA$C, axislabel=list("Dim  1  (61.81%)","Dim  2  (23.97%)") )
                              HOTcornell85 = Hotelling(cornellPCA$C, FSthreshold=0.85)
                              HotellingEllipse(HOTcornell85, colpar="black")

                         HOTcornell75 = Hotelling(cornellPCA$C, FSthreshold=0.75)
                         [1] "Outliers with components 1 and 2 ,Tē=75% :"
                              [1] "6"

                          HotellingEllipse(HOTcornell75, colpar="red")


                                        
 

Release 20.80
1) Numerical outputs in the PLS iterations are differently displayed in the one response case. The adjusted Rē  is now computed.

   > PLSL(cornell[,1:7],cornell[,8,drop=F], cexpar=0.8,bgpar="white")
   
            - Linear PLS -
   Total Variance of X = 7
   Variance of y = 1 , y0 = y

   Dimension 1

    cov(t1,y0)= 1.916  r(t1,y0)= 0.961 stdev(t1)= 1.994  stdev(y0)= 1

                    y      % of VAR
   R2 part. 0.924    92.359
  ..................
   adjusted R2 for y with dim. 1 and 7 predictors : R2_adjust= 0.791
.  .................
   % of VAR X accounted for by the current comp. = 57.361


2) PRESS plots can eventually show the threshold of acceptability for the dimension of the model.

    The rule for the dimension k, being (in the 1 response case):

                                  PRESS(k) > 1   ==>  model not better than  yhat(k) = Const = mean y 

 PRess plot  



Release 20.69
1) A new MVcut R-function to code continuous vaviables into categorical by choosing graphically the breaks points that default to quantiles.

     Heavylevel=MVcut(juice[,11,drop=F],graph=T)
     Heavy  :
     nunber of levels :  3

     Change or not that number ? (y/n)1:

                      


2) The item "Remove/Add variables" no more available in the MAPLSS menu.