Time Series Analysis with Matlab®

Version 4.6
The TSA toolbox is useful for analyzing (uni- and multivariate, stationary and non-stationary) Time Series. An Introductory tour to Time Series Analysis and the Download site can be found here.

It can be used for:

1. stochastic signal processing
2. autoregressive model identification
3. matched (inverse) filter design
4. Histogram analysis (moved to NaN-toolbox)
5. Calcution of the entropy of a timeseries
6. Non-linear analysis (3rd order statistics)
7. smoothing, prediction, filtering
8. Test for Hurwitz and Unit-Circle Polynomials
9. handles missing values (NaN's) (requires NaN-toolbox))

Several criteria (AIC, BIC, FPE, MDL, SBC, CAT, PHI) for the selection of the order of an autoregressivemodel are included. Furthermore includes the toolbox a fast version ifthe Yule-Walker method for estimating Autoregressive parameters, the AutocovarianceFunction (ACovF), Autocorrelation Function (ACF), Partial ACF (PACF),andsome other useful staff. Demo programs can be started with "demo" or "demotsa". Version 2.40 (and higher) provides fast algorithms for testing polynomials; and many functions (e.g. ACovF and the Levinson-Durbin algorithms) are implemented for multiple series.

Latest Version 4.6, 23 Sep 2019. Tested with Matlab 8.x and Octave 3.x

refer to:
Schlögl, A.; "Time Series Analysis - A toolbox for the use with Matlab", 1996-2019.
Further references on AR modeling and model order selection.

More about: the performance of used algorithms; downloading, changelog.

Back to Homepage of * * Alois* * MATLAB* * Institute for Science and Technology, Austria