conditional_SMC: Conditional SMC algorithm

Description Usage Arguments Value Author(s)

Description

Conditional SMC implementation with multinomial resampling for a generic (exchangeable) HMM

Usage

1
2
conditional_SMC(n_particles, observations, cond_trajectory, emission_density,
  transition_sam, initial_sam, ...)

Arguments

n_particles

number of particles

observations

a vector of observations on which to perform inference

cond_trajectory

a vector of states representing the trajectory on which to condition

emission_density

a function of the form f(y,x) giving the emission density of y|x

transition_sam

a function of the form f(x) sampling a new position, i.e. propagating the particle

initial_sam

a function of the form f() sampling the initial positions

...

arguments to be passed to emission_density, transition_sam and initial_sam (e.g. for Ornstein-Uhlenbeck process, provide values for sigma and delta)

Value

a list containing matrices of positions and weights (rows represent generations/time steps, columns represent individuals/particles) and an ancestry matrix (where element (i,j) is the parent in generation i of individual j in generation i+1)

Author(s)

Suzie Brown


suziebrown/miniproject1 documentation built on May 26, 2019, 9:36 a.m.