LOGLIKELIHOOD_func {seeds}

Calculates the Log Likelihood for a new sample given the current state (i.e. log[L(G|x)P(G)])

Description

Algorithm implemented according to Engelhardt et al. 2017. The function can be replaced by an user defined version if necessary.

Usage

LOGLIKELIHOOD_func(
  pars,
  Step,
  OBSERVATIONS,
  x_0,
  parameters,
  EPS_inner,
  INPUT,
  D,
  GIBBS_PAR,
  k,
  MU_JUMP,
  SIGMA_JUMP,
  eps_new,
  objectivfunc
)

Arguments

pars

sampled hidden influence for state k (w_new) at time tn+1

Step

time step of the sample algorithm corresponding to the given vector of time points

OBSERVATIONS

observed values at the given time step/point

x_0

initial values at the given time step/point

parameters

model parameters estimates

EPS_inner

current hidden inputs at time tn

INPUT

discrete input function e.g. stimuli

D

diagonal weight matrix of the current Gibbs step

GIBBS_PAR

GIBBS_PAR[["BETA"]] and GIBBS_PAR[["ALPHA"]]; prespecified or calculated vector of state weights

k

number state corresponding to the given hidden influence (w_new)

MU_JUMP

mean of the normal distributed proposal distribution

SIGMA_JUMP

variance of the normal distributed proposal distribution

eps_new

current sample vector of the hidden influences (including all states)

objectivfunc

link function to match observations with modeled states

Value

returns the log-likelihood for two given hidden inputs