A Cepstral Model for Covariate-Dependent Time Series
Bootstrap Confidence Intervals for Functional Effect Curves
Estimate Cepstral Coefficients from Periodogram
Cepstral Regression
Generate Time Series
Compute Functional Effects of Intercept and Covariates
Envelope Estimator for Log-Spectral Regression
Generate Exponential Correlation Covariance Matrix
Ordinary Least Squares Estimator for Log-Spectral Regression
Compute the Periodogram of Multivariate Time Series
Generate a Fourier Cosine Basis Matrix for Log-Spectral Modeling
Reduced-Rank Regression on Cepstral Coefficients
Fisher Scoring Algorithm For Estimating Cepstral Coefficients
Modeling associations between covariates and power spectra of replicated time series using a cepstral-based semiparametric framework. Implements a fast two-stage estimation procedure via Whittle likelihood and multivariate regression.The methodology is based on Li and Dong (2025) <doi:10.1080/10618600.2025.2473936>.