sensorMMPP: sensorMMPP

In giantoak/mmppr: Markov Modulated Poisson Process for Unsupervised Event Detection in Time Series of Counts

Description

This function provides the main MCMC inference engine

Usage

sensorMMPP(N, priors = list(aL = 1, bL = 1, aD = matrix(0, 1, 7) + 5, aH =
  matrix(0, nrow = 48, ncol = 7) + 1, z00 = 0.99 * 10000, z01 = 0.01 * 10000,
  z10 = 0.25 * 10000, z11 = 0.75 * 10000, aE = 5, bE = 1/3, MODE = 0),
  ITERS = c(50, 10), EQUIV = c(3, 3))

Arguments

N	Matrix of count data; axis 0 is the number of time intervals per day and axis 1 is the number of days in the data.
priors	List with parameter values of prior distributions
ITERS	Iteration controls: total # of iterations and # used for burn-in
EQUIV	Parameter sharing controls <- c(S1, S2): S1 <- force sharing of delta (day effect) among days, S2 <- force sharing of eta (time of day) among days, Values: 1 (all days share), 2 (weekdays/weekends), 3 (none)

giantoak/mmppr documentation built on May 17, 2019, 4:19 a.m.