R-如何从岭回归结果中提取T统计量

Question

I am making use of the ridge regression package in R, but am having trouble extracting T-statistics, p values in the the coefficients list.

Here is some sample code from here ;

data(GenCont)
mod <- linearRidge(Phenotypes ~ ., data = as.data.frame(GenCont))
summary(mod)

This returns

Call:
linearRidge(formula = Phenotypes ~ ., data = as.data.frame(GenCont))

Coefficients:
             Estimate Scaled estimate Std. Error (scaled) t value (scaled) Pr(>|t|)    
(Intercept)  1.533386              NA                  NA               NA       NA    
SNP1         0.277296        4.045409            0.266120           15.201  < 2e-16 ***
SNP2        -0.110458       -1.256154            0.216332            5.807 6.38e-09 ***

But I want to access the data in "coefficients" programmatically. I understand that if i were using lm, then something along the lines of

coef(summary(mod))

should work, but this does not seem to do what I am after. I simply get a NULL.

Any ideas on how to access this data please?

Answer 1

To get coefficient table for the situation where ridge parameters are chosen automatically you should store summary() result as some object and then access list element summaries$summary1$coefficients . Whole structure of the summary() object you can see with function str(sumar) .

sumar<-summary(mod)
sumar$summaries$summary1$coefficients
                Estimate Scaled estimate Std. Error (scaled) t value (scaled)     Pr(>|t|)
(Intercept)  1.533385893              NA                  NA               NA           NA
SNP1         0.277296215     4.045408706           0.2661197      15.20146202 0.000000e+00
SNP2        -0.110457822    -1.256153622           0.2163319       5.80660491 6.375233e-09
SNP3        -0.110457822    -1.256153622           0.2163319       5.80660491 6.375233e-09
SNP4         0.005229639     0.011635212           0.3716925       0.03130332 9.750276e-01
SNP5         0.531172545     6.323006229           0.3153685      20.04958196 0.000000e+00
SNP6        -0.119163534    -1.373227248           0.2230470       6.15667175 7.428960e-10
SNP7         0.113843942     0.113730041           0.3721807       0.30557749 7.599264e-01
SNP8        -0.099148877    -1.028580596           0.3558067       2.89084074 3.842128e-03
SNP9        -0.008320553    -0.008312229           0.3723863       0.02232152 9.821915e-01
SNP10        0.058562323     0.101128163           0.3715670       0.27216670 7.854938e-01
SNP11       -0.096526424    -1.495698673           0.3292496       4.54275034 5.552500e-06
SNP12       -0.334279101    -0.333944654           0.3722483       0.89710186 3.696646e-01

To get only t values select fourth column of this table.

> sumar$summaries$summary1$coefficients[,4]
(Intercept)        SNP1        SNP2        SNP3        SNP4        SNP5        SNP6        SNP7        SNP8        SNP9 
         NA 15.20146202  5.80660491  5.80660491  0.03130332 20.04958196  6.15667175  0.30557749  2.89084074  0.02232152 
      SNP10       SNP11       SNP12 
 0.27216670  4.54275034  0.89710186 

To access other elements produced by summary() function you just need to select appropriate list element. Structure of summary object shows which element to chose.

str(sumar$summaries$summary1)
List of 4
 $ coefficients: num [1:13, 1:5] 1.53339 0.2773 -0.11046 -0.11046 0.00523 ...
  ..- attr(*, "dimnames")=List of 2
  .. ..$ : chr [1:13] "(Intercept)" "SNP1" "SNP2" "SNP3" ...
  .. ..$ : chr [1:5] "Estimate" "Scaled estimate" "Std. Error (scaled)" "t value (scaled)" ...
 $ df          : Named num [1:3] 3.12 1.21 5.04
  ..- attr(*, "names")= chr [1:3] "model" "variance" "residual"
 $ nPCs        : int 1
 $ lambda      : num 2.21

For example to get df

sumar$summaries$summary1$df
   model variance residual 
3.120934 1.205412 5.036457 

Results can be selected also without creating new object - just use summary(mod) instead of object name sumar .

summary(mod)$summaries$summary1$df
   model variance residual 
3.120934 1.205412 5.036457