![](../logo.png)
MARSScv is a wrapper for MARSS that re-fits the model with cross validated data.
Source:R/MARSScv.R
MARSScv.Rd
MARSScv is a wrapper for MARSS that re-fits the model with cross validated data.
Arguments
- y
A n x T matrix of n time series over T time steps. Only y is required for the function. A ts object (univariate or multivariate) can be used and this will be converted to a matrix with time in the columns.
- model
Model specification using a list of parameter matrix text shortcuts or matrices. See Details and
MARSS.marxss()
for the default form. Or better yet open the Quick Start GuideRShowDoc("Quick_Start",package="MARSS")
- inits
A list with the same form as the list output by
coef(fit)
that specifies initial values for the parameters. See alsoMARSS.marxss()
.- method
Estimation method. MARSS provides an EM algorithm (
method="kem"
) (seeMARSSkem()
) and the BFGS algorithm (method="BFGS"
) (seeMARSSoptim()
). Fast TMB fitting provided by the companion package marssTMB.- form
The equation form used in the
MARSS()
call. The default is "marxss". SeeMARSS.marxss()
orMARSS.dfa()
- silent
Setting to TRUE(1) suppresses printing of full error messages, warnings, progress bars and convergence information. Setting to FALSE(0) produces error output. Setting silent=2 will produce more verbose error messages and progress information.
- control
Estimation options for the maximization algorithm. The typically used control options for method="kem" are below but see marssMLE for the full list of control options. Note many of these are not allowed if method="BFGS"; see
MARSSoptim()
for the allowed control options for this method.- fun.kf
What Kalman filter function to use. MARSS has two:
MARSSkfas()
which is based on the Kalman filter in the KFAS package based on Koopman and Durbin andMARSSkfss()
which is a native R implementation of the Kalman filter and smoother in Shumway and Stoffer. The KFAS filter is much faster.MARSSkfas()
modifies the input and output in order to output the lag-one covariance smoother needed for the EM algorithm (per page 321 in Shumway and Stoffer (2000).- fold_ids
A n x T matrix of integers, with values assigned by the user to folds. If not included, data are randomly assigned to one of 10 folds
- future_cv
Whether or not to use future cross validation (defaults to FALSE), where data up to time T-1 are used to predict data at time T. Data are held out by time slices, and the
fold_ids
argument is ignored.- n_future_cv
Number of time slices to hold out for future cross validation. Defaults to floor(n_future_cv/3). Predictions are made for the last n_future_cv time steps
- interval
uncertainty interval for prediction. Can be one of "confidence" or "prediction", and defaults to "confidence"
- ...
not used
Value
A list object, containing cv_pred
(a matrix of predictions), cv_se
(a matrix of SEs),
fold_ids
(a matrix of fold ids used as data), and df
(a dataframe containing
the original data, predictions, SEs, and folds)
Examples
# \donttest{
dat <- t(harborSealWA)
dat <- dat[2:4, ] # remove the year row
# fit a model with 1 hidden state and 3 observation time series
# cross validation here is random, 10 folds
fit <- MARSScv(dat, model = list(
Z = matrix(1, 3, 1),
R = "diagonal and equal"
))
#> Success! abstol and log-log tests passed at 23 iterations.
#> Alert: conv.test.slope.tol is 0.5.
#> Test with smaller values (<0.1) to ensure convergence.
#>
#> MARSS fit is
#> Estimation method: kem
#> Convergence test: conv.test.slope.tol = 0.5, abstol = 0.001
#> Estimation converged in 23 iterations.
#> Log-likelihood: 11.92445
#> AIC: -11.8489 AICc: -9.695058
#>
#> Estimate
#> A.SJI 0.7993
#> A.EBays 0.2778
#> R.diag 0.0234
#> U.U 0.0614
#> Q.Q 0.0095
#> x0.x0 6.1032
#> Initial states (x0) defined at t=0
#>
#> Standard errors have not been calculated.
#> Use MARSSparamCIs to compute CIs and bias estimates.
#>
#> Success! abstol and log-log tests passed at 21 iterations.
#> Alert: conv.test.slope.tol is 0.5.
#> Test with smaller values (<0.1) to ensure convergence.
#>
#> MARSS fit is
#> Estimation method: kem
#> Convergence test: conv.test.slope.tol = 0.5, abstol = 0.001
#> Estimation converged in 21 iterations.
#> Log-likelihood: 15.90228
#> AIC: -19.80456 AICc: -17.80456
#>
#> Estimate
#> A.SJI 0.7704
#> A.EBays 0.2357
#> R.diag 0.0198
#> U.U 0.0653
#> Q.Q 0.0106
#> x0.x0 6.0153
#> Initial states (x0) defined at t=0
#>
#> Standard errors have not been calculated.
#> Use MARSSparamCIs to compute CIs and bias estimates.
#>
#> Success! abstol and log-log tests passed at 25 iterations.
#> Alert: conv.test.slope.tol is 0.5.
#> Test with smaller values (<0.1) to ensure convergence.
#>
#> MARSS fit is
#> Estimation method: kem
#> Convergence test: conv.test.slope.tol = 0.5, abstol = 0.001
#> Estimation converged in 25 iterations.
#> Log-likelihood: 14.82086
#> AIC: -17.64172 AICc: -15.48787
#>
#> Estimate
#> A.SJI 0.84059
#> A.EBays 0.32619
#> R.diag 0.02188
#> U.U 0.05740
#> Q.Q 0.00638
#> x0.x0 6.18985
#> Initial states (x0) defined at t=0
#>
#> Standard errors have not been calculated.
#> Use MARSSparamCIs to compute CIs and bias estimates.
#>
#> Success! abstol and log-log tests passed at 22 iterations.
#> Alert: conv.test.slope.tol is 0.5.
#> Test with smaller values (<0.1) to ensure convergence.
#>
#> MARSS fit is
#> Estimation method: kem
#> Convergence test: conv.test.slope.tol = 0.5, abstol = 0.001
#> Estimation converged in 22 iterations.
#> Log-likelihood: 13.16553
#> AIC: -14.33107 AICc: -12.33107
#>
#> Estimate
#> A.SJI 0.77965
#> A.EBays 0.27179
#> R.diag 0.02360
#> U.U 0.06076
#> Q.Q 0.00886
#> x0.x0 6.11565
#> Initial states (x0) defined at t=0
#>
#> Standard errors have not been calculated.
#> Use MARSSparamCIs to compute CIs and bias estimates.
#>
#> Success! abstol and log-log tests passed at 23 iterations.
#> Alert: conv.test.slope.tol is 0.5.
#> Test with smaller values (<0.1) to ensure convergence.
#>
#> MARSS fit is
#> Estimation method: kem
#> Convergence test: conv.test.slope.tol = 0.5, abstol = 0.001
#> Estimation converged in 23 iterations.
#> Log-likelihood: 13.26315
#> AIC: -14.5263 AICc: -12.4263
#>
#> Estimate
#> A.SJI 0.8098
#> A.EBays 0.2662
#> R.diag 0.0215
#> U.U 0.0610
#> Q.Q 0.0112
#> x0.x0 6.0855
#> Initial states (x0) defined at t=0
#>
#> Standard errors have not been calculated.
#> Use MARSSparamCIs to compute CIs and bias estimates.
#>
#> Success! abstol and log-log tests passed at 22 iterations.
#> Alert: conv.test.slope.tol is 0.5.
#> Test with smaller values (<0.1) to ensure convergence.
#>
#> MARSS fit is
#> Estimation method: kem
#> Convergence test: conv.test.slope.tol = 0.5, abstol = 0.001
#> Estimation converged in 22 iterations.
#> Log-likelihood: 12.3994
#> AIC: -12.79881 AICc: -10.69881
#>
#> Estimate
#> A.SJI 0.78542
#> A.EBays 0.28216
#> R.diag 0.02439
#> U.U 0.06353
#> Q.Q 0.00768
#> x0.x0 6.11319
#> Initial states (x0) defined at t=0
#>
#> Standard errors have not been calculated.
#> Use MARSSparamCIs to compute CIs and bias estimates.
#>
#> Success! abstol and log-log tests passed at 22 iterations.
#> Alert: conv.test.slope.tol is 0.5.
#> Test with smaller values (<0.1) to ensure convergence.
#>
#> MARSS fit is
#> Estimation method: kem
#> Convergence test: conv.test.slope.tol = 0.5, abstol = 0.001
#> Estimation converged in 22 iterations.
#> Log-likelihood: 13.74725
#> AIC: -15.4945 AICc: -13.44572
#>
#> Estimate
#> A.SJI 0.7850
#> A.EBays 0.2763
#> R.diag 0.0218
#> U.U 0.0607
#> Q.Q 0.0102
#> x0.x0 6.1030
#> Initial states (x0) defined at t=0
#>
#> Standard errors have not been calculated.
#> Use MARSSparamCIs to compute CIs and bias estimates.
#>
#> Success! abstol and log-log tests passed at 21 iterations.
#> Alert: conv.test.slope.tol is 0.5.
#> Test with smaller values (<0.1) to ensure convergence.
#>
#> MARSS fit is
#> Estimation method: kem
#> Convergence test: conv.test.slope.tol = 0.5, abstol = 0.001
#> Estimation converged in 21 iterations.
#> Log-likelihood: 15.09518
#> AIC: -18.19037 AICc: -16.28128
#>
#> Estimate
#> A.SJI 0.79748
#> A.EBays 0.29074
#> R.diag 0.02179
#> U.U 0.06104
#> Q.Q 0.00988
#> x0.x0 6.09450
#> Initial states (x0) defined at t=0
#>
#> Standard errors have not been calculated.
#> Use MARSSparamCIs to compute CIs and bias estimates.
#>
#> Success! abstol and log-log tests passed at 24 iterations.
#> Alert: conv.test.slope.tol is 0.5.
#> Test with smaller values (<0.1) to ensure convergence.
#>
#> MARSS fit is
#> Estimation method: kem
#> Convergence test: conv.test.slope.tol = 0.5, abstol = 0.001
#> Estimation converged in 24 iterations.
#> Log-likelihood: 16.07472
#> AIC: -20.14944 AICc: -18.04944
#>
#> Estimate
#> A.SJI 0.82930
#> A.EBays 0.28020
#> R.diag 0.02024
#> U.U 0.06062
#> Q.Q 0.00757
#> x0.x0 6.09232
#> Initial states (x0) defined at t=0
#>
#> Standard errors have not been calculated.
#> Use MARSSparamCIs to compute CIs and bias estimates.
#>
#> Success! abstol and log-log tests passed at 23 iterations.
#> Alert: conv.test.slope.tol is 0.5.
#> Test with smaller values (<0.1) to ensure convergence.
#>
#> MARSS fit is
#> Estimation method: kem
#> Convergence test: conv.test.slope.tol = 0.5, abstol = 0.001
#> Estimation converged in 23 iterations.
#> Log-likelihood: 13.02546
#> AIC: -14.05091 AICc: -11.95091
#>
#> Estimate
#> A.SJI 0.7907
#> A.EBays 0.2512
#> R.diag 0.0220
#> U.U 0.0562
#> Q.Q 0.0106
#> x0.x0 6.1815
#> Initial states (x0) defined at t=0
#>
#> Standard errors have not been calculated.
#> Use MARSSparamCIs to compute CIs and bias estimates.
#>
# second, demonstrate passing in pre-specified folds
fold_ids <- matrix(
sample(1:5, size = nrow(dat) * ncol(dat), replace = TRUE),
nrow(dat), ncol(dat)
)
fit <- MARSScv(dat, model = list(
Z = matrix(1, 3, 1),
R = "diagonal and equal"
), fold_ids = fold_ids)
#> Success! abstol and log-log tests passed at 23 iterations.
#> Alert: conv.test.slope.tol is 0.5.
#> Test with smaller values (<0.1) to ensure convergence.
#>
#> MARSS fit is
#> Estimation method: kem
#> Convergence test: conv.test.slope.tol = 0.5, abstol = 0.001
#> Estimation converged in 23 iterations.
#> Log-likelihood: 13.36878
#> AIC: -14.73756 AICc: -12.33756
#>
#> Estimate
#> A.SJI 0.79203
#> A.EBays 0.23808
#> R.diag 0.02046
#> U.U 0.05710
#> Q.Q 0.00828
#> x0.x0 6.21249
#> Initial states (x0) defined at t=0
#>
#> Standard errors have not been calculated.
#> Use MARSSparamCIs to compute CIs and bias estimates.
#>
#> Success! abstol and log-log tests passed at 24 iterations.
#> Alert: conv.test.slope.tol is 0.5.
#> Test with smaller values (<0.1) to ensure convergence.
#>
#> MARSS fit is
#> Estimation method: kem
#> Convergence test: conv.test.slope.tol = 0.5, abstol = 0.001
#> Estimation converged in 24 iterations.
#> Log-likelihood: 10.93944
#> AIC: -9.878872 AICc: -7.668346
#>
#> Estimate
#> A.SJI 0.7911
#> A.EBays 0.2713
#> R.diag 0.0237
#> U.U 0.0674
#> Q.Q 0.0108
#> x0.x0 6.0100
#> Initial states (x0) defined at t=0
#>
#> Standard errors have not been calculated.
#> Use MARSSparamCIs to compute CIs and bias estimates.
#>
#> Success! abstol and log-log tests passed at 25 iterations.
#> Alert: conv.test.slope.tol is 0.5.
#> Test with smaller values (<0.1) to ensure convergence.
#>
#> MARSS fit is
#> Estimation method: kem
#> Convergence test: conv.test.slope.tol = 0.5, abstol = 0.001
#> Estimation converged in 25 iterations.
#> Log-likelihood: 11.41053
#> AIC: -10.82105 AICc: -8.550785
#>
#> Estimate
#> A.SJI 0.80315
#> A.EBays 0.26160
#> R.diag 0.02316
#> U.U 0.05675
#> Q.Q 0.00956
#> x0.x0 6.17885
#> Initial states (x0) defined at t=0
#>
#> Standard errors have not been calculated.
#> Use MARSSparamCIs to compute CIs and bias estimates.
#>
#> Success! abstol and log-log tests passed at 28 iterations.
#> Alert: conv.test.slope.tol is 0.5.
#> Test with smaller values (<0.1) to ensure convergence.
#>
#> MARSS fit is
#> Estimation method: kem
#> Convergence test: conv.test.slope.tol = 0.5, abstol = 0.001
#> Estimation converged in 28 iterations.
#> Log-likelihood: 11.88965
#> AIC: -11.7793 AICc: -9.379304
#>
#> Estimate
#> A.SJI 0.79974
#> A.EBays 0.27369
#> R.diag 0.02162
#> U.U 0.06183
#> Q.Q 0.00948
#> x0.x0 6.09562
#> Initial states (x0) defined at t=0
#>
#> Standard errors have not been calculated.
#> Use MARSSparamCIs to compute CIs and bias estimates.
#>
#> Success! abstol and log-log tests passed at 27 iterations.
#> Alert: conv.test.slope.tol is 0.5.
#> Test with smaller values (<0.1) to ensure convergence.
#>
#> MARSS fit is
#> Estimation method: kem
#> Convergence test: conv.test.slope.tol = 0.5, abstol = 0.001
#> Estimation converged in 27 iterations.
#> Log-likelihood: 11.47508
#> AIC: -10.95015 AICc: -8.325152
#>
#> Estimate
#> A.SJI 0.79874
#> A.EBays 0.31003
#> R.diag 0.02063
#> U.U 0.06061
#> Q.Q 0.00947
#> x0.x0 6.08725
#> Initial states (x0) defined at t=0
#>
#> Standard errors have not been calculated.
#> Use MARSSparamCIs to compute CIs and bias estimates.
#>
# third, illustrate future cross validation
fit <- MARSScv(dat, model = list(
Z = matrix(1, 3, 1),
R = "diagonal and equal"
), future_cv = TRUE, n_future_cv = 5)
#> Success! abstol and log-log tests passed at 21 iterations.
#> Alert: conv.test.slope.tol is 0.5.
#> Test with smaller values (<0.1) to ensure convergence.
#>
#> MARSS fit is
#> Estimation method: kem
#> Convergence test: conv.test.slope.tol = 0.5, abstol = 0.001
#> Estimation converged in 21 iterations.
#> Log-likelihood: 14.91458
#> AIC: -17.82916 AICc: -15.87567
#>
#> Estimate
#> A.SJI 0.80143
#> A.EBays 0.28885
#> R.diag 0.02367
#> U.U 0.06778
#> Q.Q 0.00652
#> x0.x0 6.10448
#> Initial states (x0) defined at t=0
#>
#> Standard errors have not been calculated.
#> Use MARSSparamCIs to compute CIs and bias estimates.
#>
#> Success! abstol and log-log tests passed at 27 iterations.
#> Alert: conv.test.slope.tol is 0.5.
#> Test with smaller values (<0.1) to ensure convergence.
#>
#> MARSS fit is
#> Estimation method: kem
#> Convergence test: conv.test.slope.tol = 0.5, abstol = 0.001
#> Estimation converged in 27 iterations.
#> Log-likelihood: 16.33737
#> AIC: -20.67474 AICc: -18.57474
#>
#> Estimate
#> A.SJI 0.79451
#> A.EBays 0.30766
#> R.diag 0.02326
#> U.U 0.07594
#> Q.Q 0.00345
#> x0.x0 6.11259
#> Initial states (x0) defined at t=0
#>
#> Standard errors have not been calculated.
#> Use MARSSparamCIs to compute CIs and bias estimates.
#>
#> Success! abstol and log-log tests passed at 32 iterations.
#> Alert: conv.test.slope.tol is 0.5.
#> Test with smaller values (<0.1) to ensure convergence.
#>
#> MARSS fit is
#> Estimation method: kem
#> Convergence test: conv.test.slope.tol = 0.5, abstol = 0.001
#> Estimation converged in 32 iterations.
#> Log-likelihood: 17.44618
#> AIC: -22.89236 AICc: -20.62209
#>
#> Estimate
#> A.SJI 0.80612
#> A.EBays 0.33869
#> R.diag 0.02147
#> U.U 0.08159
#> Q.Q 0.00269
#> x0.x0 6.09125
#> Initial states (x0) defined at t=0
#>
#> Standard errors have not been calculated.
#> Use MARSSparamCIs to compute CIs and bias estimates.
#>
#> Success! abstol and log-log tests passed at 32 iterations.
#> Alert: conv.test.slope.tol is 0.5.
#> Test with smaller values (<0.1) to ensure convergence.
#>
#> MARSS fit is
#> Estimation method: kem
#> Convergence test: conv.test.slope.tol = 0.5, abstol = 0.001
#> Estimation converged in 32 iterations.
#> Log-likelihood: 15.81173
#> AIC: -19.62347 AICc: -17.15288
#>
#> Estimate
#> A.SJI 0.79297
#> A.EBays 0.34634
#> R.diag 0.02173
#> U.U 0.08298
#> Q.Q 0.00297
#> x0.x0 6.08537
#> Initial states (x0) defined at t=0
#>
#> Standard errors have not been calculated.
#> Use MARSSparamCIs to compute CIs and bias estimates.
#>
#> Success! abstol and log-log tests passed at 29 iterations.
#> Alert: conv.test.slope.tol is 0.5.
#> Test with smaller values (<0.1) to ensure convergence.
#>
#> MARSS fit is
#> Estimation method: kem
#> Convergence test: conv.test.slope.tol = 0.5, abstol = 0.001
#> Estimation converged in 29 iterations.
#> Log-likelihood: 16.2755
#> AIC: -20.55101 AICc: -17.84133
#>
#> Estimate
#> A.SJI 0.79445
#> A.EBays 0.38057
#> R.diag 0.01893
#> U.U 0.08406
#> Q.Q 0.00393
#> x0.x0 6.05363
#> Initial states (x0) defined at t=0
#>
#> Standard errors have not been calculated.
#> Use MARSSparamCIs to compute CIs and bias estimates.
#>
# }