[R] multi-response models in the pls package

Ben Wasserman ben.a.wasserman at gmail.com
Tue Feb 23 18:21:16 CET 2010


Hello List,

I am using the function plsr (from package pls) to fit a model with 41
independent variables and several (8ish) response variables.  I'm
using Leave-one-out cross validation.

I would like to know how to find out the total variation in Y
explained by a model with a given number of components.
summary(myplsr) reports the total variation in X explained, and
variation in each Y individually explained (ex below)

TRAINING: % variance explained
                                                      1 comps  2 comps
 3 comps  4 comps
X                                                         16.77603
23.029    29.54    35.18
Number.of.successful.strikes                 10.57655   21.108    23.69    26.86
First.successful.strike.top                       0.98759    1.775
15.71    16.57
First.successful.strike.bottom                12.00975   31.560
37.25    41.88
Number.of.successful.strikes.bottom        0.06298    3.920    10.15    10.19
Number.of.successful.strikes.top           10.57665   21.108    23.69    26.86
Average.interval.successful.strike.top      0.34717    2.367    14.05    16.99
Average.interval.successful.strike.bottom19.65802   44.453    53.76    61.38
Average.interval.successful.strike           15.07655   15.274    15.27    21.50

Thanks for your help,
Ben



More information about the R-help mailing list