QoLSEM: Quality of Life analysis using a structural equation model...

In AntoineBbi/QoLSEM: Quality of Life analysis using a Structural Equation Model (QoLSEM)

Quality of Life analysis using a structural equation model (QoLSEM)

Description

This function fits a structural equation model for the analysis of the Quality of life data from EORTC questionnaires. The estimation is achieved through the EM algorithm. More details can be found in the article cited reference section.

Usage

QoLSEM(y, X1, X2, Ty = NULL, T1 = NULL, T2 = NULL,
  convergence = 0.001, nb_it = 500, trace = FALSE)

Arguments

y	a numeriacl vector of observations of the response variable
X1	matrix of a first block of data associated with the first factor
X2	matrix of a second block of data associated with the second factor
Ty	matrix/vecteur of covariable(s) associated with Y
T1	matrix/vecteur covariable(s) associated with X1
T2	matrix/vecteur covariable(s) associated with X2
convergence	a numerial scalar denoting the criteria of convergence to stop the iteration; =0.001 by default
nb_it	a numerical scalar denoting the maximal number of iterations of the EM algorithm; =500 by default
trace	if TRUE, the function shows for each iteration the convergence of algorithm (default=FALSE)

Value

A list with the following elements:

Factors

matrix of size (nx2) of the predicted factors. The first (reps. second) column is associated with the first (resp. second) factor

C

Vector of 2 scalars corresponding to the estimated parameters c1 and c2

A1

vector of parameters associated with the factor in the first variable block

A2

vector of parameters associated with the factor in the second variable block

D

vector of parameters associated with the covariates in the structural equation

D1

matrix of parameters associated with the covariates in block 1

D2

matrix of parameters associated with the covariates in block 2

estimate_sigma2

vector of variance parameters

Convergence

argument diff

Iteration

Number of iterations until the convergence

Author(s)

Antoine Barbieri, Myriam Tami

References

Barbieri A, Tami M, Bry X, Azria D, Gourgou S, Mollevi C, Lavergne C. (2018) EM algorithm estimation of a structural equation model for the longitudinal study of the quality of life. Statistics in Medicine. 37(6):1031-1046.

See Also

generation.QoLSEM

Examples

## generation of a dataset
test <- generation.QoLSEM(N=1,
                          I=150,
                          c=as.matrix(c(2,-2)),
                          a1=matrix(c(1:7),nrow=7),
                          a2=matrix(c(1:12),nrow=12),
                          d=c(80),
                          D1=matrix(seq(2,14,2)+70,ncol=7),
                          D2=matrix(seq(2,24,2)+20,ncol=12),
                          sigma.y=10,
                          sigma.X1=10,
                          sigma.X2=10)

## Estimation of the model
simu <- QoLSEM(test$y,
               test$X1,
               test$X2,
               Ty=NULL,
               T1=NULL,
               T2=NULL,
               convergence=0.001,
               nb_it=500,
               trace=FALSE)

## Predicted subject-speficic factors
simu$Factors

## Estimation parameters associated with the intercept of the block 1
simu$D1

AntoineBbi/QoLSEM documentation built on May 10, 2022, 6:07 a.m.