
600 Current Trends in Bayesian Methodology with Applications
1950 1960 1970 1980 1990 2000 2010
−2 0 2 4
year
U
[t]
+ β
FIGURE 28.7
Posterior predictive check. The full time series is simulated from the poster ior
predictive distribution 5000 times, and 95% po intwise posterior predictive
intervals are given by the dashed lines. Gray circles represent the o bserved
G(Y ) = ω
T
˜
Y . The observed data are consistent with the po sterior predictive
distribution.
all three proxies are observed. The simulated series are c onsistent with the
observed data.
Since solar physicis ts are concerned with predicting the timing of solar
cycles, we can obtain fitted values