Rfssa: An R Package for Functional Singular Spectrum Analysis

Authors

• Hossein Haghbin, Persian Gulf University
• Jordan Trinka, Marquette University
• Mehdi Maadooliat, Marquette UniversityFollow

Document Type

Article

Publication Date

6-2024

Publisher

The R Foundation

Source Publication

The R Journal

Source ISSN

2073-4859

Original Item ID

DOI: 10.32614/RJ-2024-019

Abstract

Functional Singular Spectrum Analysis (FSSA) is a non-parametric approach for analyzing Functional Time Series (FTS) and Multivariate FTS (MFTS) data. This paper introduces Rfssa, an R package that addresses implementing FSSA for FTS and MFTS data types. Rfssa provides a flexible container, the funts class, for FTS/MFTS data observed on one-dimensional or multi-dimensional domains. It accepts arbitrary basis systems and offers powerful graphical tools for visualizing time-varying features and pattern changes. The package incorporates two forecasting algorithms for FTS data. Developed using object-oriented programming and Rcpp/RcppArmadillo, Rfssa ensures computational efficiency. The paper covers theoretical background, technical details, usage examples, and highlights potential applications of Rfssa.

Comments

The R Journal, Vol. 16, No. 2 (June 2024): 82-99. DOI.

Recommended Citation

Haghbin, Hossein; Trinka, Jordan; and Maadooliat, Mehdi, "Rfssa: An R Package for Functional Singular Spectrum Analysis" (2024). Mathematical and Statistical Science Faculty Research and Publications. 175.
https://epublications.marquette.edu/math_fac/175