fix stochastic volatility examples

Author: Ben Goodrich

misc/moving-avg/arma11-sim.R misc/moving-avg/arma11.data.Rmisc/moving-avg/arma11-sim.R renamed to misc/moving-avg/arma11.data.R

@@ -12,5 +12,5 @@ y[1] <- err[1] + mu + phi * mu;
 for (t in 2:T)
   y[t] <- err[t] + (mu + phi * y[t-1] + theta * err[t-1]);
 
-library('rstan')
+#library('rstan')
 # fit <- stan('arma11.stan', data=list(T=T, y=y), iter=2000, chains=4);

misc/moving-avg/arma11_alt.data.R

@@ -0,0 +1,16 @@
+mu <- -1.25;
+sigma <- 0.75;
+theta <- 0.5;
+phi <- 0.2;
+
+T <- 1000;
+
+err <- rnorm(T,0,sigma);
+nu <- rep(0,T);
+y <- rep(0,T);
+y[1] <- err[1] + mu + phi * mu;
+for (t in 2:T)
+  y[t] <- err[t] + (mu + phi * y[t-1] + theta * err[t-1]);
+
+#library('rstan')
+# fit <- stan('arma11.stan', data=list(T=T, y=y), iter=2000, chains=4);

misc/moving-avg/ma2-sim.R misc/moving-avg/ma2.data.Rmisc/moving-avg/ma2-sim.R renamed to misc/moving-avg/ma2.data.R

@@ -20,5 +20,5 @@ for (t in 3:T) {
   epsilon[t] <- y[t] - predictor;
 }
 
-library('rstan')
+#library('rstan')
 # fit <- stan('ma2.stan', data=list(T=T, y=y), iter=500, chains=2, init=0);

misc/moving-avg/stochastic-volatility-optimized.data.R

@@ -0,0 +1,15 @@
+
+phi <- 0.95;
+sigma <- 0.25;
+beta <- 0.6;
+mu <- 2 * log(beta);
+
+T <- 500;
+
+h <- rep(NA,T);
+h[1] <- rnorm(1, mu, sigma / sqrt(1 - phi * phi));
+for (t in 2:T)
+  h[t] <- rnorm(1, mu + phi * (h[t-1] - mu), sigma);
+y <- rep(NA,T);
+for (t in 1:T)
+  y[t] <- rnorm(1, 0, exp(h[t] / 2));

misc/moving-avg/stochastic-volatility-sim.R

@@ -1,20 +1,5 @@
-
-phi <- 0.95;
-sigma <- 0.25;
-beta <- 0.6;
-mu <- 2 * log(beta);
-
-T <- 500;
-
-h <- rep(NA,T);
-h[1] <- rnorm(1, mu, sigma / sqrt(1 - phi * phi));
-for (t in 2:T)
-  h[t] <- rnorm(1, mu + phi * (h[t-1] - mu), sigma);
-y <- rep(NA,T);
-for (t in 1:T)
-  y[t] <- rnorm(1, 0, exp(h[t] / 2));
-
+source("stochastic-volatility.data.R")
 library("rstan");
 start_time <- proc.time();
-fit <- stan(file="stochastic-volatility.stan", data=list(T=T,y=y), iter=10000, chains=4, init=0, fit=fit);
+fit <- stan(file="stochastic-volatility.stan", data=list(T=T,y=y), iter=10000, chains=4, init=0);
 elapsed_time <- proc.time() - start_time;

misc/moving-avg/stochastic-volatility.data.R

@@ -0,0 +1,15 @@
+
+phi <- 0.95;
+sigma <- 0.25;
+beta <- 0.6;
+mu <- 2 * log(beta);
+
+T <- 500;
+
+h <- rep(NA,T);
+h[1] <- rnorm(1, mu, sigma / sqrt(1 - phi * phi));
+for (t in 2:T)
+  h[t] <- rnorm(1, mu + phi * (h[t-1] - mu), sigma);
+y <- rep(NA,T);
+for (t in 1:T)
+  y[t] <- rnorm(1, 0, exp(h[t] / 2));