mvlsboot: mvlsboot

In helpstatanalysis/mvls: Missing Values in Longitudinal Studies

Description

Takes a longitudinal dataset and impute missing value with a machine learning-based method many time.

Usage

mvlsboot(data, d = 0.1, method = "k", cluster = 6, nstart = 20,
  pre.imp = F, imp.method = "mean", boot = 10)

Arguments

data	A dataset (more than two longitudinal mesurements)
d	It is the percentage of change between two-sided mesurements to consider it bigger, smaller or the same, It useful to built the var.matrix
method	It represent the type of machine learning algorithm. 'k' for k-mean and 'h' for hierical
cluster	It's the number of cluster. Default setting it's 6. It depends on number of longitudinal mesurements. It could be use mvls.print to decide best cluster number.
nstart	It is the nstart setting of function k-mean. Defualt it'20. Not requested for 'h' method.
pre.imp	TRUE/FALSE (default F). It permit to pre-impute data to built the vari.matrix, It could be reduce cluster with only missing value.
imp.method	It's the type of pre-imputation. Defaul it's 'mean', but there is also 'locf' possibility
boot	Impute many times to reduce error from cluster imputation. 'low' for 5 imputation, 'medium' for 10 imputation and 'high' for 15 imputation.

Value

$data It's the data-set with imputation.

$sd.2 It contain the sd for each data imputed at multiple imputation method. Different from sd.1.

$db.boot It's a list of all imputed dataset (n=boot).

Author(s)

Lorenzo Querci <[e-mail]>

helpstatanalysis/mvls documentation built on May 13, 2019, 3 a.m.