Time Series Library - tslib

The library contains a set of functions for manipulation and processing of 3D-time-series arrays

Description

The library contains a set of functions for handling 3d arrays originated by manipulation with MIMO systems. The functions convert data from 3D arrays to time-series 2D matrices and backward, and enables some algebraic operations between them. The library is in one directory named tslib.

The directory contains also a short description of the library in a pdf-file.

The library is complemented by a demo, in which some of the functions are used to calculate responses of a linear two-degrees-of-freedom system to an arbitrary excitation.

• MATLAB release: MATLAB 7.2 (R2006a)
• Other requirements: For running demoTSL file, the following functions are necessary: inp.m - FEX ID 9033 gig.m - FEX ID 9035

[Dimitri Shvorob]

I think that the title is misleading: the functions can be used to work with multivariate time series, but these are not date/time functions per se. I also wish the PDF offered more details on the functions: surprisingly, a chapter named 'An example' shows no Matlab code. Personally, I have not had trouble working with 3D array - how hard is it to code a loop, when 'x(i,j,:)' can't be used? - but other people may find your code useful, if you just explain upfront what it does. (For example, 'tsarr2mtx - ts array to ts matrix conversion' - what is that?). How about an HTML sample session?

[Dibakar Datta]

excillent

[Peter -]

The key-issue in calculations with 3d matrices is the fact that for-loops are slow in matlab, compared to matrix operations. When operating MIMO sound-samples for example the number of iterations of a for-loop is like 44100 per second of the sample. This library uses for-loops and therefore doesn't really solve the problem

[carlos lopez]