MLRMPA (version 1.0)

stepMLRMPA: Multiple Linear Regression Model Population Analysis

Description

Following the variance and correletion selection, the function stepMLRMPA is to perform the clustering procedures, random sampling, building a model, validation the model and printing the model and statistic parameters to the specified files.

Usage

stepMLRMPA(tree, Clusters, N, op1, op2,tr.tst)

Arguments

tree

a hierarchical tree of variables resulting from hclustvar into several clusters by specifying the desired number of clusters

Clusters

an integer scalar with the desired number of clusters

N

an integer scalar with the desired number of sampling and modeling times

op1

a txt file used for save the statistical parameters

op2

a txt file used for save the model

tr.tst

a list calculated by VarCor function

Value

Clusterth_size

the number of variables in each cluster

Details

The specific procedure can be seen in the figure 1.

References

M.Chavent, V. Kuentz, B.Liquet, L. Saracco, J. Stat. Softw. 2012, 50, 1-16.

See Also

hclustvar VarCor

Examples

##  'var.lower' and 'var.upper' are the variance selection range 
var.lower<-0.05
var.upper<-0.07

##  'xy.cor' is the correlation selection range
xy.cor<-0.3

## 'tst' is the number of test set
tst<-c(72:101)
data(activity)
data(descriptor)
data(deleted_descriptor)
tr.tst<-VarCor(tst,activity,descriptor,deleted_descriptor,var.lower,var.upper,xy.cor)

## the variance distribution of the original dataset
plotvar(tr.tst[[3]],tr.tst[[1]])

## the variance distribution of the dataset with variance and correlation selection
plotvar(tr.tst[[4]],tr.tst[[1]])

## the correlation distribution of the original dataset
plotcor(tr.tst[[4]],tr.tst[[1]])

## the correlation distribution of the dataset with variance and correlation selection
plotcor(tr.tst[[4]],tr.tst[[1]])

xtr<-as.data.frame(tr.tst[[4]])
ytr<-as.data.frame(tr.tst[[1]])
xtst<-as.data.frame(tr.tst[[5]])
ytst<-as.data.frame(tr.tst[[2]])

## variables clustering
tree<-hclustvar(xtr)

Clusterth.SIze2<-stepMLRMPA(tree,2,5,op1="statistic_parameters02.txt",op2="model02.txt",tr.tst)
Clusterth.SIze3<-stepMLRMPA(tree,3,5,op1="statistic_parameters03.txt",op2="model03.txt",tr.tst)