Value Members

1.  object DeterminePosterior extends App with TestModel

   Run the PMMH algorithm over the first 400 simulated observations 1.

   Run the PMMH algorithm over the first 400 simulated observations 1. Read the data in from a file 2. group the first 400 observations into a Seq[Data] 3. use mapConcat to get a tuple containing the MCMC chain identifier and a Seq[Data] 4. Use mapAsync to run two chains in parallel 5. Write the chains to file in JSON format

2.  object Filtering extends App with TestModel

   Filter using the same parameters that were used to simulate the model: 1.

   Filter using the same parameters that were used to simulate the model: 1. Read in data simulated from the model, using DataFromFile 2. Define the particle filter and the start time for the filter 3. Run the filter using the same parameters we simulated the model with 4. Save the output to a file asynchronously

3.  object Interpolate extends App with TestModel
4.  object OnlineFiltering extends App with TestModel

   Once the posterior distribution of the state and parameters, p(x, theta | y) has been determined, filtering can be performed online 1.

   Once the posterior distribution of the state and parameters, p(x, theta | y) has been determined, filtering can be performed online 1. Read in the test data, dropping the first 4000 elements which are used to determine the posterior 2. Set up the filter 3. Read in the posterior distribution from a JSON file 4. Run several filters each starting with a sample (x, theta) from the joint posterior of the state and parameters 5. Write the results of the filter to a file

5.  object PilotRun extends App with TestModel

   Perform a pilot run, by running the particle filter over the data multiple times to determine the variance of the estimate of the log-likelihood using different amount of particles.

   Perform a pilot run, by running the particle filter over the data multiple times to determine the variance of the estimate of the log-likelihood using different amount of particles. In order to run the PMMH algorithm quickly, a low number of particles is required, but an accurate estimate of the likelihood is also required. A rule of thumb for the variance of the estimated log-likelihood is 1.0

6.  object SimOrnstein extends App with TestModel
7.  object SimulateModel extends App with TestModel

   Simulate Data from the Test model and write it to CSV and JSON files asynchronously