readData: Read Data and Edit Rules

In EditImputeCont: Simultaneous Edit-Imputation for Continuous Microdata

Description

add description about this function

Usage

readData(Y.original, ratio = NULL, range = NULL, balance = NULL, eps.bal = 0.6) 

Arguments

Y.original	original dataset of (n, p) dimension with missing and edit-failing values where n is the number of records and p is the number of variables.
ratio	ratio edit.
range	range restriction.
balance	balance edit.
eps.bal	threshold for balance edit. Defaults to 0.6.

Details

Y.original has n records and p variables. The variable names (column names) of Y.original are used to specify ratio edits.

The edit rules are either imported from text files or written by editrules package's syntax.

A balance edit is considered as two inequality constraints with the threshold, i.e., ‘A = B’ is converted to ‘-eps.bal < A - B < eps.bal’ before computation.

For accurate computation, nested balances are written as ‘total variable = sum of component variables’. For example, it is recommended to replace ‘X1 = X2 + X3’ and ‘X3 = X4 + X5’ with ‘X1 = X2 + X4 + X5’ and ‘X3 = X4 + X5’ so that ‘X3’ does not appear both sides of the balance edits.

Value

readData returns an EditIn.data object which consists of

• Y.input: input dataset which replaces NA in Y.original with -999 and zero values with 0.01.

• Edit.editmatrix: editmatrix of edit rules. It can be used for functions of editrules package.

• Edit.matrix:matrix of edit rules.

• Bound.LU: range restrictions. For variable X whose range is not specified in range, the default values are set as max( 0.1min(X), 1e-5 ) for the lower bound and 10max(x) for the upper bound.

• ratio: ratio edits.

• n.balance: number of balance edit, i.e., the row number of balance.

• FaultyRecordID: record IDs of Y.orig whose values violate edit rules.

References

Hang J. Kim, Lawrence H. Cox, Alan F. Karr, Jerome P. Reiter and Quanli Wang (2015). "Simultaneous Edit-Imputation for Continuous Microdata", Journal of the American Statistical Association, DOI: 10.1080/01621459.2015.1040881.

Edwin de Jonge and Mark van der Loo (2013). editrules: R package for parsing, applying, and manipulating data cleaning rules. R package version 2.7.2.

See Also

editmatrix

Examples

### option 1. import from text files ###

data(NestedEx)
 
D_obs1 = NestedEx$D.obs
Ratio1 = NestedEx$Ratio.edit
Range1 = NestedEx$Range.edit
Balance1 = NestedEx$Balance.edit

data1 = readData(Y.original=D_obs1, ratio=Ratio1, range=Range1, 
balance=Balance1, eps.bal=0.6)

# print(data1$Edit.editmatrix)
# plot(data1$Edit.editmatrix)	  ## function of 'editrules' package

### option 2. Using the syntax of R package 'editrules' ###

data(NestedEx) ; D_obs2 = NestedEx$D.obs

Ratio2 <- editmatrix(c(
 "X1 <= 1096.63*X5", "X1 <= 2980.96*X7", "X1 <= 148.41*X8", "X1 <= 7.39*X9",
 "X5 <= 0.37*X1", "X5 <= 54.60*X7", "X5 <= 2.72*X8", "X5 <= 0.14*X9",
 "X7 <= 0.14*X1", "X7 <= 1.65*X5", "X7 <= 7.39*X8", "X7 <= 0.05*X9",
 "X8 <= 1.65*X1", "X8 <= 54.60*X5", "X8 <= 403.43*X7", "X8 <= 1.65*X9",
 "X9 <= 20.09*X1", "X9 <= 403.43*X5", "X9 <= 13359.73*X7", "X9 <= 148.41*X8"
))
Range2 <- editmatrix(c(
 "X1 >= 2", "X2 <= 1.2e+06", "X11 >= 0.002", "X11 <= 1.2e+04"
))
Balance2 <- editmatrix(c(
 "X1 == X2+X3+X4", "X5 == X6 + 0.4*X10 + 0.6*X11", "X7 == 0.4*X10 + 0.6*X11"
))

data2 = readData(D_obs2, Ratio2, Range2, Balance2)

# print(data2$Edit.editmatrix)
  # Note: data2 is equivalent to data1
# plot(data2$Edit.editmatrix)	  ## function of 'editrules' package

EditImputeCont documentation built on March 26, 2020, 7:15 p.m.