fitMisRepARMA {MisRepARMA}

Fit ARMA model to misreported time series data

Description

Fits an ARMA model to misreported time series data.

Usage

  fitMisRepARMA(y, tol, B, p_AR, q_MA, covars=NULL, misReport="U", ...)

Arguments

y

a numeric vector or time series giving the original data.

tol

tolerance limit to stop the iterative algorithm.

B

the number of bootstrap series to compute.

p_AR

order of the AR part.

q_MA

order of the MA part.

covars

matrix of explanatory variables. Its default value is NULL.

misReport

direction of misreporting issue. Its default value is U for underreported data, can also take the value O for overreported data.

...

additional arguments to pass to tsboot, for instance those regarding parallelization.

Details

The model based resampling scheme with B bootstrap resamples is computed. This

Value

An object of class fitMisRepARMA with the following elements is returned:

• data: The original time series.

• t0: The results of applying statistic to the original series.

• t: Estimates on each replicated time series.

• call: The original call to tsboot.

Author(s)

David Moriña, Amanda Fernández-Fontelo, Alejandra Cabaña, Pedro Puig

References

Davison, A.C. and Hinkley, D.V. (1997) Bootstrap Methods and Their Application. Cambridge University Press.

Kunsch, H.R. (1989) The jackknife and the bootstrap for general stationary observations. Annals of Statistics, 17, 1217–1241.

Moriña, D., Fernández-Fontelo, A., Cabaña, A., Puig, P. (2021): New statistical model for misreported data with application to current public health challenges. arXiv preprint (https://arxiv.org/pdf/2003.09202.pdf)

Politis, D.N. and Romano, J.P. (1994) The stationary bootstrap. Journal of the American Statistical Association, 89, 1303–1313.

See Also

MisRepARMA-package, reconstruct

Examples

### Simulate underreported time series data
set.seed(12345)
x <- arima.sim(model=list(ar=0.4), n=50)
ind <- rbinom(50, 1, 0.6)
y <- ifelse(ind==0, x, x*0.3)
mod <- fitMisRepARMA(y, 1e-6, 3, 0.05, 1, 0, covars=NULL, misReport="U")