README.md

In shimlab/HMTree: Hidden Markov tree prior for analysis of multiple functional data

This repository contains an R package for implementing Hidden Markov tree prior.

To install the HMTree package first you need to install devtools,roxygen2 and Rcpp

install.packages("devtools")
install.packages("Rcpp")
library("devtools")
library("Rcpp")
install_github("shimlab/HMTree")

Running HMTree

To simulate data under either Poisson or Normal model, you can run 'simu.curves' to simulated data. To get help:

library(HMTree)
?simu.curves

For example, you can simulate 10 curve of length 1024 under Poisson model or Normal model using:

curve.length = 1024
model.mode = 'Poisson'
num.samples=10
set.seed(666)
res.pois = simu.curves(curve.length=curve.length, model.mode=model.mode, num.samples=num.samples)

curve.length = 1024
model.mode = 'Normal'
num.samples=10
normal.sigma=0.5
set.seed(666)
res.normal = simu.curves(curve.length=curve.length, model.mode=model.mode, num.samples=num.samples, normal.sigma=normal.sigma)

Or you can use you own data, by using "HMTree.parameter" function to generate your own data object: Note: in user input data matrix, number of rows should be the number of samples and number of column should be the length.

mydata.normal<-res.normal$data
mydata.poisson<-res.poisson$data
HMTree.parameter(mydata.normal,"Normal")
HMTree.parameter(mydata,"Poisson")

Next step is denoise. For one dimensional data, either Normal or Poisson, you can run 'HMTree.denoise_1D'. To get help:

?HMTree.denoise_1D

For example, after simulating data, you can get denoised objects by:

res.denoised.pois<-HMTree.denoise_1D(res.pois)
res.denoised.normal<-HMTree.denoise_1D(res.normal)

Meaning of each object: alpha_vec wavelet_coef denoised_wavelet_coef denoised param_signa_sqr param_marginal_probs param_transition_probs post_probs post_trans_probs

shimlab/HMTree documentation built on May 29, 2019, 9:25 p.m.