Modern Statistical Methods for HCI by Judy Robertson & Maurits Kaptein

Author:Judy Robertson & Maurits Kaptein
Language: eng
Format: epub
Publisher: Springer International Publishing, Cham

Our two distributions for and are declared within the JAGS model. JAGS stands for Just Another Gibbs Sampler and it is a program for analyzing Bayesian models using MCMC sampling. We first supply parameters and their assigned prior beliefs (which are uniform in this case) to our model. We also need to define the likelihood. Binomial distribution is our choice for this problem which is a discrete distribution using parameters as the probability rate (probability of using a cell phone) and the total number of sequences. From a programming perspective this may appear a bit peculiar, however, models do not have to operate sequentially or have functions and distributions being sequentially defined. As long as all variables, distributions and parameters are all accounted for, the model will produce results. In this case, s1, s2, n1, n2 are known discrete variables and and are unknown albeit with priors defined. Finally, we can also calculate the difference between and called (in the R code defined as delta).

After setting up our model we can proceed by setting the parameters for MCMC using the command jag.model and then utilize coda.samples to generate posterior samples based on the parameters of interest. The process simulates all variables for our model however the sampled chains that are returned are only those that interest us. These are declared as a list.

Download

Copyright Disclaimer:
This site does not store any files on its server. We only index and link to content provided by other sites. Please contact the content providers to delete copyright contents if any and email us, we'll remove relevant links or contents immediately.