calculateChangePoints: calculateChangePoints

In kojikoji/nbcf: Change point detection for time course multi negative binomial observation

Description

Function for detecting change points and changed variates

Usage

calculateChangePoints(count.mat, t.vec, mean.bias.vec, alpha = 1,
  beta = 1, r = 30, lambda = 0.01, p.res = 1000,
  count.res = 1000, t.res = 100, sim.iter = 100, method = "map",
  map.num = 1)

Arguments

count.mat	dgCmatrix, Count matrix. Each row represent each variate. Each column represent each observation
t.vec	Numeric vector, Observation time
mean.bias.vec	Numeric vector, Relative mean bias for each observation
r	Numeric, Size parameters of negative binomial distribution.
lambda	Numeric, Prior probability the change point occur each step
p.res	Integer, Approximation resolution of parameter p
count.res	Integer, Approximation resolution of count.
t.res	Integer, Approximation resolution of t.
sim.iter	Integer, Iteration of perfect simulation. See perfectSimulation
method	Specify the method to calculate change points. "map" or "sim"
map.num	Number of change points. This works only if you specify method as "map"

Details

This function is responsible for main calculation of this package. This calculate simulated change points for each variates, and provide them a data frame

Value

change.point.df, Data frame, This contains simulated change points for each variates. This is composed of columns of changed variates, change point coordinates.

Author(s)

Yasuhiro Kojima

kojikoji/nbcf documentation built on June 16, 2019, 6:08 p.m.