Define a model to use throughout the examples in this file
Sample from the joint posterior of the state and parameters p(x, theta | y) Serialize this to JSON using Spray JSON Write as invalid JSON, by converting each element of the sequence to JSON and writing them on a new line of the output file This is the same as Twitters streaming API
Determine the parameters of the seasonal model using PMMH *
Run the filter over the last 100 elements of the simulted data using samples from the joint-posterior of the state and parameters, p(x, theta | y)
Filter from the start of the series with the parameters used to simulate the model *
Determine the full joint posterior of the state and the parameters of the Linear Model then serialise the results to JSON, so it can be read in and used for the online filtering
Run 100 particle filters by sampling 100 times from p(x, theta | y), each filter has M particles
Once a pilot run of the PMMH algorithm has been completed, the covariance of the posterior can be used as the covariance of the proposal distribution
Perform a one step forecast of the data
Perform a one step forecast on the poisson data, using unseen test data, Sampling from the joint posterior of the parameters and the state p(x, theta | y)
Perform a pilot run of the PMMH algorithm, to determine the optimum number of particles to use in the particle filter
Determine how many particles are required to run the MCMC
Serve the data as a stream of JSON
Determine the appropriate amount of particles in the particle filter for the seasonal model *
Simulate an SDE
Simulate a poisson model, with seasonal rate parameter