Package 'HhP'

Title: Hierarchical Heterogeneity Analysis via Penalization
Description: In medical research, supervised heterogeneity analysis has important implications. Assume that there are two types of features. Using both types of features, our goal is to conduct the first supervised heterogeneity analysis that satisfies a hierarchical structure. That is, the first type of features defines a rough structure, and the second type defines a nested and more refined structure. A penalization approach is developed, which has been motivated by but differs significantly from penalized fusion and sparse group penalization. Reference: Ren, M., Zhang, Q., Zhang, S., Zhong, T., Huang, J. & Ma, S. (2022). "Hierarchical cancer heterogeneity analysis based on histopathological imaging features". Biometrics, <doi:10.1111/biom.13426>.
Authors: Mingyang Ren [aut, cre] , Qingzhao Zhang [aut], Sanguo Zhang [aut], Tingyan Zhong [aut], Jian Huang [aut], Shuangge Ma [aut]
Maintainer: Mingyang Ren <[email protected]>
License: GPL-2
Version: 1.0.0
Built: 2024-12-25 06:37:54 UTC
Source: CRAN

Help Index

Hierarchical Heterogeneity Regression Analysis.

Description

The main function for Transfer learning for tensor graphical models.

Usage

evaluation.sum(n,q,p,admmres, abic.n, admmres2, Beta0, bic.var)

Arguments

n

The sample size.
q

The dimension of type 1 features.
p

The dimension of type 2 features.
admmres

The results corresponding to lambda1.
abic.n

The BIC values.
admmres2

The results corresponding to lambda1.
Beta0

The true values of beta.
bic.var

The BIC values.

Value

A result list including: evaluating indicator

Some example data

Description

Some example data

Format

A list.

Source

Simulated data

Examples

data(example.data.GGM)

Some example data

Description

Some example data

Format

A list.

Source

Simulated data

Examples

data(example.data.reg)

Hierarchical Heterogeneity Regression Analysis.

Description

The main function for Transfer learning for tensor graphical models.

Usage

gen_int_beta(n, p, q, whole.data, subgroup=c(2,4),
                    ridge = FALSE, gr.init=10, lambda.min=0.0001)

Arguments

n

The sample size.
p

The dimension of type 2 features.
q

The dimension of type 1 features.
whole.data

The input data analyzed (a list including the response and design matrix).
subgroup

When using fmrs to generate initial value, the initial value parameter of fmrs is given. Randomly divide this number of groups into several groups.
ridge

The logical variable, whether or not to yield initial values using ridge regression.
gr.init

The subgroup number of initial values using ridge regression.
lambda.min

The tuning parameter using ridge regression, the default is 0.0001.

Value

A result list.

Author(s)

Mingyang Ren, Qingzhao Zhang, Sanguo Zhang, Tingyan Zhong, Jian Huang, Shuangge Ma. Maintainer: Mingyang Ren [email protected].

References

Mingyang Ren, Qingzhao Zhang, Sanguo Zhang, Tingyan Zhong, Jian Huang, Shuangge Ma. 2022. Hierarchical Cancer Heterogeneity Analysis Based On Histopathological Imaging Features. Biometrics, <DOI: 10.1111/biom.13544>.

Examples

library(HhP)
library(Matrix)
library(MASS)
library(fmrs)
data(example.data.reg)
n = example.data.reg$n
q = example.data.reg$q
p = example.data.reg$p

beta.init.list  =  gen_int_beta(n, p, q, example.data.reg)
beta.init  =  beta.init.list$beta.init
lambda  =  genelambda.obo()
result  =  HhP.reg(lambda, example.data.reg, n, q, p, beta.init)
index.list  =  evaluation.sum(n,q,p, result$admmres, result$abic.n,
               result$admmres2, example.data.reg$Beta0, result$bic.var)
index.list$err.s

Generate tuning parameters

Description

Generating a sequence of the tuning parameters (lambda1 and lambda2).

Usage

genelambda.obo(nlambda1=20,lambda1_max=0.5,lambda1_min=0.1,
                      nlambda2=5,lambda2_max=1.5,lambda2_min=0.1)

Arguments

nlambda1

The numbers of lambda 1.
lambda1_max

The maximum values of lambda 1.
lambda1_min

The minimum values of lambda 1.
nlambda2

The numbers of lambda 2.
lambda2_max

The maximum values of lambda 2.
lambda2_min

The minimum values of lambda 2.

Value

A sequence of the tuning parameters (lambda1, lambda2, and lambda3).

Author(s)

Mingyang Ren

Examples

lambda <- genelambda.obo()
lambda

Hierarchical Heterogeneity Regression Analysis.

Description

The main function for Transfer learning for tensor graphical models.

Usage

HhP.reg(lambda, whole.data, n, q, p, beta.init,
               merge.all=FALSE, trace=FALSE, selection.sub=FALSE)

Arguments

lambda

The sequences of the tuning parameters (lambda1 and lambda2).
whole.data

The input data analyzed (a list including the response and design matrix).
n

The sample size.
q

The dimension of type 1 features.
p

The dimension of type 2 features.
beta.init

The Initial values of regression coefficients.
merge.all

the logical variable, the default is F.
trace

the logical variable, whether or not to output the number of identified subgroups during the search for parameters.
selection.sub

the logical variable, the default is F.

Value

A result list.

Author(s)

Mingyang Ren, Qingzhao Zhang, Sanguo Zhang, Tingyan Zhong, Jian Huang, Shuangge Ma. Maintainer: Mingyang Ren [email protected].

References

Mingyang Ren, Qingzhao Zhang, Sanguo Zhang, Tingyan Zhong, Jian Huang, Shuangge Ma. 2022. Hierarchical Cancer Heterogeneity Analysis Based On Histopathological Imaging Features. Biometrics, <DOI: 10.1111/biom.13544>.

Examples

library(HhP)
library(Matrix)
library(MASS)
library(fmrs)
data(example.data.reg)
n = example.data.reg$n
q = example.data.reg$q
p = example.data.reg$p

beta.init.list  =  gen_int_beta(n, p, q, example.data.reg)
beta.init  =  beta.init.list$beta.init
lambda  =  genelambda.obo()
result  =  HhP.reg(lambda, example.data.reg, n, q, p, beta.init)
index.list  =  evaluation.sum(n,q,p, result$admmres, result$abic.n,
               result$admmres2, example.data.reg$Beta0, result$bic.var)
index.list$err.s