## Two approaches to set up an order-two Markov model ## Works x[1] ~ dnNormal(0,1) delta ~ dnNormal(0,sqrt(1/3)) x[2] := x[1] + delta for (i in 1:98) { x[i+2] ~ dnNormal(2 * x[i+1] - x[i],1) } print("Take 1, calculating log-probability of x[1]") print(x[1].lnProbability()) clear() ## Non-centered parameterization. Doesn't work (Rev will need to be force-closed) x[1] ~ dnNormal(0,1) delta ~ dnNormal(0,sqrt(1/3)) x[2] := x[1] + delta for (i in 1:98) { dx[i] ~ dnNormal(0,1) x[i+2] := dx[i] - x[i] + 2 * x[i+1] # # Use below print statement to watch how time to compute probability slows as model is built print(x[1].lnProbability()) } print("Take 2, calculating log-probability of x[1]") print(x[1].lnProbability()) moves[1] = mvSlide(x[1]) mymodel = model(x[1]) # mymodel.graph("temp.dot",verbose=true,bg="honeydew2") monitors[1] = mnScreen() mymcmc = mcmc(mymodel, moves, monitors) mymcmc.run(10) q()