longitudinalData (version 0.6.4)

imputation: ~ Function: imputation ~

Description

imputation is a function that offer different methods to impute missing value of a LongData.

Usage

imputation(object, method, partition)

Arguments

object

[LongData] or [matrix] : longitudinal data to impute

method

[character]: Name of the imputation method (see detail)

partition

[Partition] : for some imputation technique (like "copyMean"), a specific partition is needed. See detail.

Value

• A matrix with no missing values.

item

• LOCB (Last Occurence Carried Backward)
• Missing at the end:
• linearInterpolation
• Missing at start / at the end:
• linearInterpolation2
• Missing at start:
• Missing at the end:
• linearInterpolation3
• Missing at start:
• Missing at the end:
• copyMean:

itemize

• Missing in the middle:

code

Partition

Details

imputation is a function that impute missing value of a LongData. Several imputation methods are available. For each method, the imputation has to deal with three kind of missing value : at start of the trajectorie (first values are missing), at the end (last values are missing) or in the middle (the missing value have surround by non-missing value). Here is a description of each methods (for all of them, an example is provided in the Examples section):

• LOCF (Last Occurence Carried Forward)

{

• Missing in the middle / at the end:

{ the previous non-missing value is dipplicated forward.} Missing at start:{ the first non-missing occurence is dupplicated backward (LOCB).} }

See Also

LongData, Partition, criterion

Examples

##################
### Preparation of the data
timeV <- 1:19
trajMissing <- longData(
   matrix(c(NA,NA ,2 ,3   ,NA,5 ,5.5,5.8,6 ,NA ,NA,6.5 ,7.5 ,NA ,NA ,NA ,4 ,NA,NA,
            2 ,0.5,-2,-2.6,2 ,1 ,1.5,0  ,-2,1.2,1 ,-3.5,-4.9,0.7,1.2,2.5,-1,-1, 1),2,byrow=TRUE),
   id=1:2,time=timeV,varName="V"
)

plot(timeV,trajMissing["traj"][1,],col=2,type="o",lwd=3,ylim=c(-2,10),
  xlab="Trajectorie to impute",ylab="")
par(ask=TRUE)

##################
### LOCF
trajImp <- imputation(trajMissing,method="LOCF")
plot(timeV,trajImp["traj"][1,],type="o",ylim=c(-2,10),ylab="",xlab="LOCF")
lines(timeV,trajMissing["traj"][1,],col=2,type="o",lwd=3)

### LOCB
trajImp <- imputation(trajMissing,method="LOCB")
plot(timeV,trajImp["traj"][1,],type="o",ylim=c(-2,10),ylab="",xlab="LOCB")
lines(timeV,trajMissing["traj"][1,],col=2,type="o",lwd=3)

### linearInterpolation
trajImp <- imputation(trajMissing,method="linearInterpolation")
plot(timeV,trajImp["traj"][1,],type="o",ylim=c(-2,10),ylab="",xlab="linearInterpolation")
lines(timeV,trajMissing["traj"][1,],col=2,type="o",lwd=3)

### linearInterpolation2
trajImp <- imputation(trajMissing,method="linearInterpolation2")
plot(timeV,trajImp["traj"][1,],type="o",ylim=c(-2,10),ylab="",xlab="linearInterpolation2")
lines(timeV,trajMissing["traj"][1,],col=2,type="o",lwd=3)

### linearInterpolation3
trajImp <- imputation(trajMissing,method="linearInterpolation3")
plot(timeV,trajImp["traj"][1,],type="o",ylim=c(-2,10),ylab="",xlab="linearInterpolation3")
lines(timeV,trajMissing["traj"][1,],col=2,type="o",lwd=3)

### copyMean
meanTraj <- apply(trajMissing["traj"],2,meanNA)
plot(timeV,trajMissing["traj"][1,],type="o",ylim=c(-2,10),col=2,lwd=3,ylab="",
  xlab="LongData to impute and its model (in green)")
lines(timeV,meanTraj,col=3,type="b",lwd=3)

trajImp <- imputation(trajMissing,method="copyMean",partition=partition(nbCluster=1,clusters=c(1,1)))
plot(timeV,trajImp["traj"][1,],type="o",ylim=c(-2,10),ylab="",xlab="copyMean")
lines(timeV,trajMissing["traj"][1,],col=2,type="o",lwd=3)
lines(timeV,meanTraj,col=3,type="o",lwd=3)


### copyMean with several clusters
trajMiss <- longData(
   matrix(c(2,3,NA,0, NA,4,0,NA, 4,NA,-1,0),4),
   id=1:4,time=1:3
)
part <- partition(nbCluster=2,clusters=c(1,1,2,2))
imputation(trajMiss,method="copyMean",partition=part)
par(ask=FALSE)