generateHMM: generateHMM

In jiangrongo/HMM:

Description

generateHMM generates multiple trajectures of observed chain X according to the hidden Markov model.

Usage

generateHMM(num, n = 100, trans = matrix(c(0, 3, 7, 5, 2, 3, 5, 5, 0),
  3, 3, byrow = TRUE)/10, u = c(-5, 0, 5), sig = c(1, 1, 1),
  pi0 = c(1, 1, 1)/3)

Arguments

num	number of Markov chains
n	length of Markov chain (assume all chains have equal length)
trans	a matrix of transition probability
u	emission prob: mean. For the rest of arguments, the dimension must match with trans.
sig	emission prob: standard deviation.
pi0	initial probability.

Details

Assumes hidden Markov model, the emmision probability follows normal distribution.

• transition probability is specified by user via trans. Default value: [0, 3, 7; 5, 2, 3; 5, 5, 0]/10

• emission prob: mean is specified by user via u. It must match the dimension of trans. Defualt vaule u = [0,5,-5].

• emission prob: sd is specified by user via sig. It must match the dimension of trans. Defualt vaule sig=[1,1,1].

• initial probability. Default value pi0=[1,1,1]/3

Value

A list containing:

• matrix of hidden state, Z. Each column is a Markov chain.

• matrix of observed states, X. (each column corresponds to a column of Z)

• pi0, the initial probability

• trans

• u, sig

Examples

set.seed(1221)
df <- generateHMM(num=2,n=10); df

jiangrongo/HMM documentation built on May 19, 2019, 9:38 p.m.