Homepage > MCMC
Markov Chain Monte Carlo
  • Concept of Distribution
  • Monte Carlo Simulation
  • Markov Chain
  • Metropolis Algorithm
    Gibbs Sampler
  • Dynamics of Gibbs Models
  • Gibbs Distribution
  • Ising and Potts Models
    Bayesian Models
  • Joint Distribution
  • Frequentist Approach
  • Bayesian Approach
  • Dirichlet Distribution
  • Mixture Model
  • MCMC scheme
    Data Mining and MCMC
  • Relative Report Rate
  • Graphical Model
  • MCMC scheme
    Explore MCMC with R
  • Normal Mixture
  • Searching for Maxima
  • Bernoulli Trials
  • Random Beta Walk
  • Metropolis Algorithm
  • Potts Model

    • MCMC scheme

    We can define the density $ \pi$ over the graphs of drug-adverse reaction relationship by

    $\displaystyle \pi(G) \propto \prod_{(A,B)} \Phi(m_{A,B}, E_{A,B}) $

    where the product is over all the association candidate $ (A,B)$ indicated by the graph $ G$ . Here we use the Metropolis algorithm to update the graph $ G$ by randomly adding or deleting an edge between drug and adverse reaction (i.e., a drug-adverse reaction relationship).

    Dataset. The present dataset freq.csv consists of 13343 rows. The first 22 columns of 0's and 1's indicates the subset $ A$ of the reported medical products (drug), and the next 27 columns of 0's and 1's indicates the subset $ B$ of the adversed reactions, and the last column shows the frequency $ N_{A,B}$ of event. The names of drug and the medical terms describing the adverse events can be found in the separate files drug.csv and reac.csv.

    Original database. The dataset freq.csv is extracted from the US database from January 2004 to March 2005 available by FDA; see the link to AERS.

    Explore it. Download aers.r together with datasets freq.csv, drug.csv and reac.csv. 

    > source("aers.r")
> aers()

    © TTU Mathematics