adaptspec: Adaptive Spectral Estimation for Non-stationary Time Series
In BayesSpec: Bayesian Spectral Analysis Techniques

Description Usage Arguments Value Author(s) References Examples

Methodology for analyzing possibly non-stationary time series by adaptively dividing the time series into an unknown but finite number of segments and estimating the corresponding local spectra by smoothing splines.

adaptspec(nloop, nwarmup, nexp_max, x,
   tmin, sigmasqalpha, tau_prior_a, tau_prior_b,
   tau_up_limit, prob_mm1, step_size_max,
   var_inflate, nbasis, nfreq_hat, plotting)

`nloop`	The total number of MCMC iterations
`nwarmup`	The number of burn-in iterations
`nexp_max`	The maximum number of segments allowed
`x`	The data, a univariate time series, not a time series object
`tmin`	The minimum number of observations per segment. An optional argument defaulted to tmin = 40.
`sigmasqalpha`	An optional argument defaulted to sigmasqalpha = 100.
`tau_prior_a`	An optional argurment defaulted to tau_prior_a = -1.
`tau_prior_b`	An optional argurment defaulted to tau_prior_b = 0.
`tau_up_limit`	An optional argurment defaulted to tau_up_limit = 10000.
`prob_mm1`	An optional argurment defaulted to prob_mm1 = 0.8.
`step_size_max`	An optional argurment defaulted to step_size_max = 10.
`var_inflate`	An optional argurment defaulted to var_inflate = 1.
`nbasis`	An optional argurment defaulted to nbasis = 7.
`nfreq_hat`	An optional argurment defaulted to nfreq_hat = 50.
`plotting`	An optional argument for displaying output plots defaulted to FALSE. When set to TRUE, this displays the spectral and parition points.

xi The partition points

log_spec_hat Estimates of the log spectra for all segments

nexp_curr The number of segments in each iteration.

Rosen, O., Wood, S. and Stoffer, D.

Rosen, O., Wood, S. and Stoffer, D. (2012). AdaptSPEC: Adaptive Spectral Estimation for Nonstationary Time Series. J. of the American Statistical Association, 107, 1575-1589

#Running adaptspec with the simulated_piecewise data.
data(simulated_piecewise)
model1 <- adaptspec(nloop = 80, nwarmup = 20,
   nexp_max = 5, x = simulated_piecewise[1:100])
str(model1)
summary(model1$nexp_curr)
plot(model1$nexp_curr)