mice (version 2.2)

mice.impute.fastpmm: Imputation by fast predictive mean matching

Description

Imputes univariate missing data using fast predictive mean matching

Usage

mice.impute.fastpmm(y, ry, x, donors = 5, type = 1, ridge = 1e-05, ...)

Arguments

y

Numeric vector with incomplete data

ry

Response pattern of y (TRUE=observed, FALSE=missing)

x

Design matrix with length(y) rows and p columns containing complete covariates.

donors

Size of the set of potential donors from which a random draw is made. The default is donors = 5.

type

Type of matching distance. The default choice type = 1 calculates the distance between the predicted value of yobs and the drawn values of ymis. Other choices are type = 0 (distance between predi

ridge

The ridge penalty applied in .norm.draw() to prevent problems with multicollinearity. The default is ridge = 1e-05, which means that 0.01 percent of the diagonal is added to the cross-product. Larger ridges may result in

...

Other named arguments.

Value

• Numeric vector of length sum(!ry) with imputations

Details

Imputation of y by predictive mean matching, based on Rubin (1987, p. 168, formulas a and b). The procedure is as follows:

1. Estimate beta and sigma by linear regression
2. Draw beta* and sigma* from the proper posterior
3. Compute predicted values foryobsbeta andymisbeta*
4. For eachymis, find the observation with closest predicted value, and take its observed value inyas the imputation.
5. If there is more than one candidate, make a random draw among them. Note: The matching is done on predictedy, NOT on observedy.

References

Little, R.J.A. (1988), Missing data adjustments in large surveys (with discussion), Journal of Business Economics and Statistics, 6, 287--301.

Rubin, D.B. (1987). Multiple imputation for nonresponse in surveys. New York: Wiley.