naive_bayes.jags

model {
    theta[1:Ntopics] ~ ddirch( onesTopics )
    for ( i in 1:Ntopics ) {
      phi[i] ~ ddirch( onesVocab )
    }
    
    for ( i in 1 : Ndocs ) {
      z[i] ~ dcat( theta[1:Ntopics] )
    }
    for ( i in 1 : Nwords ) {
      w[i] ~ dcat( phi[z[doc[i]], ] ) 
    }
}