Package 'robustETM'

Robust Exponential Tilt Mixture Model

Description

The package robustETM consists of the functions to perform pseudolikelihood based EM test for homogeneity in generalized exponential tilt mixture models.

Details

Package:	robustETM
Type:	robustETM
Version:	1.0
Date:	2016-03-27
License: GPL>=2

Testing for homogeneity in generalized exponential titl mixture model

Motivated by analyses of DNA methylation data, we propose a semiparametric mixture model, namely the generalized exponential tilt mixture model, to account for heterogeneity between differentially methylated and non-differentially methylated subjects in the cancer group, and capture the differences in higher order moments (e.g. mean and variance) between subjects in cancer and normal groups. A pairwise pseudolikelihood is constructed to eliminate the unknown nuisance function. To circumvent boundary and non-identifiability problems as in parametric mixture models, we modify the pseudolikelihood by adding a penalty function. In addition, test with simple asymptotic distribution has computational advantages over permutational test for high-dimensional genetic and epigenetic data. We propose a pseudolikelihood based expectation-maximization test, and show the proposed test follows a simple chi-squared limiting distribution.

The methods contains in function sim are:

• The proposed PLEMT test (pseudolikelihood based EM test)

• The t-test

• The modified empirical likelihood ratio test

• The empirical likelihood ratio test

• The logistic regression test

• The Wilcoxon test

• The F test

• The KS test

Author(s)

Chuan Hong
Yong Chen
Yang Ning
Hao Wu Maintainer: Chuan Hong <[email protected]>

References

Hong, C., Chen Y., Ning Y., Wang S., Wu H. and Carroll R.J. (2016). PLEMT: A novel pseudolikelihood based EM test for homogeneity in generalized exponential tilt mixture model (in preparation).

---

Tests under comparision for testing for homogeneity in generalized exponential tilt mixture models

Description

The function conducts the pseudolikelihood based EM test for homogeneity in generalized exponential tilt mixture models

Usage

sim(itr, K, cc, i.n, isetting, lambda, distn)

Arguments

itr	random seed
K	Number of grid values for proportion parameter lambda
cc	Tuning parameter C for penalty function
isetting	Type I error or power scenarios I II and III for simulation study
lambda	Proportion parameter lambda
i.n	Sample size setting
distn	Distribution

Value

mplrt_EM.TS	Test statistic for the proposed PLEMT test
qin.TS	Test statistic for empirical likelihood ratio test
liu.TS	Test statistic for modified empirical likelihood ratio test
t.TS	Test statistic for t-test
wilcox.p	p-value for wilcoxon test
logist.TS	Test statistic for logistic regression test

Author(s)

Chuan Hong, Yong Chen, Yang Ning, Hao Wu

References

Hong, C., Chen Y., Ning Y., Wang S., Wu H. and Carroll R.J. (2016). PLEMT: A novel pseudolikelihood based EM test for homogeneity in generalized exponential tilt mixture model (in preparation).

Examples

# not run
#myresult=sim(itr=1234, K=10, cc=20, i.n=2, isetting=1, lambda=0.3, distn="norm")

---