comp_error: Compute imputation error

In udellgroup/mixedgcImp: Missing value imputation for mixed data through Gaussian copula

Compute imputation error

Description

Compute the imputation error on entries which are missing in xobs but observed in xtrue

Usage

cal_mae(xhat, xobs, xtrue, round = FALSE)

cal_rmse(xhat, xobs, xtrue, relative = TRUE)

cal_smae(xhat, xobs, xtrue, round = FALSE, reduce = TRUE, verbose = TRUE)

cal_misclass(xhat, xobs, xtrue)

Arguments

xhat	imputed data matrix or vector
xobs	incomplete observed data matrix or vector
xtrue	complete true data matrix or vector
round	whether to round the values of xhat to observed integers in xobs
relative	Only used for cal_rmse. If TRUE, normalize the computed error by the Frobenius norm of xtrue
reduce	Only used for cal_smae. Return the average error if TRUE else the whole error vector
verbose	Only used for cal_smae. If TRUE, throw out a warning when perfect baseline imputation appears

Value

Computed imputation error

Functions

• cal_mae: Compute the mean absolute error

• cal_rmse: Compute the root mean squared error

• cal_smae: Compute the MAE scaled by median imputation MAE (at each column)

• cal_misclass: Compute the mis-classification rate for categorical data

udellgroup/mixedgcImp documentation built on Jan. 25, 2023, 7:55 p.m.