Suppressed tabular data: Yhat from X and Z

Implementation of equation 21 in the paper.

Usage

Z2Yhat(z, x, digits = 9)

Arguments

z: Z as a matrix
x: X as a matrix
digits: When non-NULL, output values close to whole numbers will be rounded using digits as input to RoundWhole.

Value

Yhat as a matrix

Details

Generalized inverse is computed by ginv. In practise, the computations can be speeded up using reduced versions of X and Z. See ReduceX.

Author

Øyvind Langsrud

Examples

# Same data as in the paper
z <- RegSDCdata("sec7z")
x <- RegSDCdata("sec7x")
Z2Yhat(z, x)
#>                freq
#> row1_col1  4.173913
#> row2_col1  4.521739
#> row3_col1 12.000000
#> row4_col1 13.304348
#> row1_col2  9.826087
#> row2_col2 10.173913
#> row3_col2 22.000000
#> row4_col2 19.000000
#> row1_col3 32.000000
#> row2_col3  8.304348
#> row3_col3  6.695652
#> row4_col3 16.000000
#> row1_col4 30.000000
#> row2_col4  8.000000
#> row3_col4 -2.695652
#> row4_col4  7.695652

# With y known, yHat can be computed in other ways
y <- RegSDCdata("sec7y")  # Now z is t(x) %*% y 
fitted(lm(y ~ x - 1))
#> row1_col1 row2_col1 row3_col1 row4_col1 row1_col2 row2_col2 row3_col2 row4_col2 
#>  4.173913  4.521739 12.000000 13.304348  9.826087 10.173913 22.000000 19.000000 
#> row1_col3 row2_col3 row3_col3 row4_col3 row1_col4 row2_col4 row3_col4 row4_col4 
#> 32.000000  8.304348  6.695652 16.000000 30.000000  8.000000 -2.695652  7.695652 
IpsoExtra(y, x, FALSE, resScale = 0)
#>                freq
#> row1_col1  4.173913
#> row2_col1  4.521739
#> row3_col1 12.000000
#> row4_col1 13.304348
#> row1_col2  9.826087
#> row2_col2 10.173913
#> row3_col2 22.000000
#> row4_col2 19.000000
#> row1_col3 32.000000
#> row2_col3  8.304348
#> row3_col3  6.695652
#> row4_col3 16.000000
#> row1_col4 30.000000
#> row2_col4  8.000000
#> row3_col4 -2.695652
#> row4_col4  7.695652