Compack (version 0.1.0)

cv.compCL: Cross-validation for compCL.

Description

k-fold cross-validation for compCL; produce a plot and return optimal values of lam.

Usage

cv.compCL(y, Z, Zc = NULL, intercept = FALSE, lam = NULL,
          nfolds = 10, foldid, trim = 0, keep = FALSE, ...)

Arguments

y

response vector with length n.

Z

z matrix as in compCL.

Zc

Zc matrix as in compCL. Default is NULL.

intercept

whether to include an intercept. Default is FALSE.

lam

a user supplied lambda sequence. If lam is provided as a scaler and nlam\(>1\), lam sequence is created starting from lam. To run a single value of lam, set nlam\(=1\). The program will sort user-defined lambda sequence in decreasing order.

nfolds

number of folds, default is 10. The smallest allowable value is nfolds=3.

foldid

an optional vector of values between 1 and the sample size n, providing the fold assignments. If supplied, nfold can be missing.

trim

percentage to be trimmed off the prediction errors from either side; default is 0.

keep

If keep=TRUE, fitted models in cross validation are reported. Default is keep=FALSE.

...

other arguments that can be passed to compCL.

Value

an object of S3 class "cv.compCL" is returned, which is a list constaining:

compCL.fit

a fitted compCL object for the full data.

lam

the sequence of lam.

Ftrim

a list of cross-validation results without trimming:

• cvm the mean cross-validated error - a vector of length length(lam).

• cvsd standard error of cvm.

• cvupper upper curve = cvm+cvsd.

• cvlo lower curve = cvm-cvsd.

• lam.min the optimal value of lam that gives minimum cross validation error.

• lam.1se the largest value of lam such that the error is within 1 standard error of the minimum cvm.

Ttrim

a list of cross-validation result with trim*100%, The structure is the same as that for Ftrim.

foldid

the values of foldid.

Details

cross-validation and fit full data with selected model.

References

Lin, W., Shi, P., Peng, R. and Li, H. (2014) Variable selection in regression with compositional covariates, https://academic.oup.com/biomet/article/101/4/785/1775476. Biometrika 101 785-979

See Also

compCL and cv.compCL, and coef, predict and plot methods for "cv.compCL" object.

Examples

# NOT RUN {
p = 30
n = 50
beta = c(1, -0.8, 0.6, 0, 0, -1.5, -0.5, 1.2)
beta = c(beta, rep(0, times = p - length(beta)))
Comp_data = comp_Model(n = n, p = p, beta = beta, intercept = FALSE)
cvm1 <- cv.compCL(y = Comp_data$y, Z = Comp_data$X.comp,
                  Zc = Comp_data$Zc, intercept = Comp_data$intercept)

plot(cvm1)
coef(cvm1)
## selection by "lam.min" criterion
which(abs(coef(cvm1, s = "lam.min")[1:p]) > 0)
## selection by "lam.1se" criterion
which(abs(coef(cvm1, s= "lam.1se")[1:p]) > 0)

Comp_data2 = comp_Model(n = 30, p = p, beta = Comp_data$beta, intercept = FALSE)
y_hat = predict(cvm1, Znew = Comp_data2$X.comp, Zcnew = Comp_data2$Zc)
plot(Comp_data2$y, y_hat,
     xlab = "Observed response", ylab = "Predicted response")

# }