diabetes: Blood and other measurements in diabetics

In biglars: Scalable Least-Angle Regression and Lasso

Description

The diabetes1 data frame has 442 rows and 1 columns. These are the data used in the Efron et al "Least Angle Regression" paper.

The diabetes data frame has 442 rows and 3 matrices, containing predictors, response, and interactions.

Format

diabetes1 is a data frame with 442 observations on the following 11 variables.

age

a numeric vector

sex

a numeric vector

bmi

a numeric vector

map

a numeric vector

tc

a numeric vector

ldl

a numeric vector

hdl

a numeric vector

tch

a numeric vector

ltg

a numeric vector

glu

a numeric vector

y

a numeric vector

In the sex variable, 1 indicates female and 2 male.

diabetes is a data frame containing the following objects:

x

a matrix with 10 columns–the first 10 columns from diabetes1, standardized

y

a numeric vector

x2

a matrix with 64 columns–main effects and second-order interactions

Details

The x matrix is standardized to have unit L2 norm in each column and zero mean. The matrix x2 consists of x plus second-order powers and interactions, also standardized.

Source

http://www-stat.stanford.edu/~hastie/Papers/LARS/LeastAngle_2002.ps

References

B. Efron and T. Hastie (2003), "LARS software for R and Splus", http://www-stat.stanford.edu/~hastie/Papers/LARS

B. Efron, T. Hastie, I. Johnstone and R. Tibshirani (2004), "Least Angle Regression" (with discussion), Annals of Statistics 32, 407-499.

Examples

data(diabetes)

stepFit <- biglars.fit(diabetes$x, diabetes$y, type = "stepwise")
stepFitBlocked <- biglars.fit(diabetes$x, diabetes$y, type = "stepwise",
                           blockSize = 50)

lassoFit <- biglars.fit(diabetes$x, diabetes$y)
lassoFitBlocked <- biglars.fit(diabetes$x, diabetes$y, blockSize = 34)

biglars documentation built on May 2, 2019, 3:08 a.m.