Become an expert in R — Interactive courses, Cheat Sheets, certificates and more!
Get Started for Free

makeX

convert a data frame to a data matrix with one-hot encoding


Description

Converts a data frame to a data matrix suitable for input to glmnet. Factors are converted to dummy matrices via "one-hot" encoding. Options deal with missing values and sparsity.

Usage

makeX(train, test = NULL, na.impute = FALSE, sparse = FALSE, ...)

Arguments

train

Required argument. A dataframe consisting of vectors, matrices and factors

test

Optional argument. A dataframe matching 'train' for use as testing data

na.impute

Logical, default FALSE. If TRUE, missing values for any column in the resultant 'x' matrix are replaced by the means of the nonmissing values derived from 'train'

sparse

Logical, default FALSE. If TRUE then the returned matrice(s) are converted to matrices of class "CsparseMatrix". Useful if some factors have a large number of levels, resulting in very big matrices, mostly zero

...

additional arguments, currently unused

Details

The main function is to convert factors to dummy matrices via "one-hot" encoding. Having the 'train' and 'test' data present is useful if some factor levels are missing in either. Since a factor with k levels leads to a submatrix with 1/k entries zero, with large k the sparse=TRUE option can be helpful; a large matrix will be returned, but stored in sparse matrix format. Finally, the function can deal with missing data. The current version has the option to replace missing observations with the mean from the training data. For dummy submatrices, these are the mean proportions at each level.

Value

If only 'train' was provided, the function returns a matrix 'x'. If missing values were imputed, this matrix has an attribute containing its column means (before imputation). If 'test' was provided as well, a list with two components is returned: 'x' and 'xtest'.

Author(s)

Trevor Hastie
Maintainer: Trevor Hastie hastie@stanford.edu

See Also

glmnet

Examples

set.seed(101)
### Single data frame
X = matrix(rnorm(20), 10, 2)
X3 = sample(letters[1:3], 10, replace = TRUE)
X4 = sample(LETTERS[1:3], 10, replace = TRUE)
df = data.frame(X, X3, X4)
makeX(df)
makeX(df, sparse = TRUE)

### Single data freame with missing values
Xn = X
Xn[3, 1] = NA
Xn[5, 2] = NA
X3n = X3
X3n[6] = NA
X4n = X4
X4n[9] = NA
dfn = data.frame(Xn, X3n, X4n)

makeX(dfn)
makeX(dfn, sparse = TRUE)
makeX(dfn, na.impute = TRUE)
makeX(dfn, na.impute = TRUE, sparse = TRUE)

### Test data as well
X = matrix(rnorm(10), 5, 2)
X3 = sample(letters[1:3], 5, replace = TRUE)
X4 = sample(LETTERS[1:3], 5, replace = TRUE)
dft = data.frame(X, X3, X4)

makeX(df, dft)
makeX(df, dft, sparse = TRUE)

### Missing data in test as well
Xn = X
Xn[3, 1] = NA
Xn[5, 2] = NA
X3n = X3
X3n[1] = NA
X4n = X4
X4n[2] = NA
dftn = data.frame(Xn, X3n, X4n)

makeX(dfn, dftn)
makeX(dfn, dftn, sparse = TRUE)
makeX(dfn, dftn, na.impute = TRUE)
makeX(dfn, dftn, sparse = TRUE, na.impute = TRUE)

glmnet

Lasso and Elastic-Net Regularized Generalized Linear Models

v4.1-1
GPL-2
Authors
Jerome Friedman [aut], Trevor Hastie [aut, cre], Rob Tibshirani [aut], Balasubramanian Narasimhan [aut], Kenneth Tay [aut], Noah Simon [aut], Junyang Qian [ctb]
Initial release
2021-02-17

We don't support your browser anymore

Please choose more modern alternatives, such as Google Chrome or Mozilla Firefox.