HOSA - Higher Order Spectral Analysis Toolbox

Spectral and polyspectral analysis, and time-frequency distributions.

Description

There is much more information in a stochastic non-Gaussian or deterministic signal than is conveyed by its autocorrelation and power
spectrum. Higher-order spectra which are defined in terms of the higher-order moments or cumulants of a signal, contain this additional information. The Higher-Order Spectral Analysis (HOSA) Toolbox provides comprehensive higher-order spectral analysis capabilities for signal processing applications. The toolbox is an excellent resource for the advanced researcher and the practicing engineer, as well as the novice
student who wants to learn about concepts and algorithms in statistical signal processing.

The HOSA Toolbox is a collection of M-files that implement a variety of advanced signal processing algorithms for the estimation of cross- and auto-cumulants (including correlations), spectra and olyspectra,bispectrum, and bicoherence, and omputation of time-frequency
distributions. Based on these, algorithms for parametric and non-parametric blind system identification, time-delay estimation, harmonic retrieval, phase-coupling, direction of arrival estimation, parameter estimation of Volterra (non-linear) models, and adaptive linear prediction are implemented. Also included are algorithms for testing of Gaussianity and Linearity of a time series. A full tutorial and demo set are included in the toolbox.

• Required Products: Signal Processing Toolbox
• MATLAB release: MATLAB 6.5 (R13)

[John Brown]

 

[- -]

Is this the same as TMW's HOSA Toolbox?

[George Barakos]

Pity it is not part of the main MatLab release!

[Dr. Cornelis Los]

This the best non-wavelet, higher-order time-frequency analysis of time series and the only one freely available. Use it closely together with the excellent book for the best results and for easy and accurate interpretation.

[D. Gansawat]

 

[Usama Zaher]

Excellent application

If there are more examples it would be nice.Especially if it is concerned with some application of sensor validation

[guojun yang]

 

[x yg]

 

[Ma Yongfeng]

It is good and helpful.

[extra ll]

 

[dylan mckay]

 

[Mokrane Malek]

 

[Moh Al-jebori]

 

[w w]

very good

[Dale Clifford]

This IS a matlab toolbox and installs the same even as far as the icon in the view window of Matlab. Previously not available as a student edition, but now free.This toolbox provides much information on signals, especially if the researcher is looking for non-linear dynamics. Works well when used along the side of wavelets.

[qiang xueming]

excellent

[Jaime Aviles]

 

[Norbert Marwan]

Why this toolbox is not provided as a zip-package or Matlab-installation file (eg. created with Makeinstall). Moreover, all filenames are upperletter. The exe-package as well as the upperletter filenames make it hard for a Unix user to use this toolbox.

[Brian Powell]

Excellent toolbox. The only negative is that it is distributed as an EXE. If you distribute as a tar, it should work everywhere. After unpacking on a borrowed windows box, I use it on Solaris and MacOSX.

[Camilo Londoño]

Nice work!!!!

[Dr. Abdul-Bary Sulaiman]

A very useful toolbox, but i\I surprised that it is only found in V5.3 Matlab and has been cancelled in the uper versions !!!

[wang shenghui]

ok

[zheng qing]

 

[xnkj xnkj]

hao dong dong

[Marcus Kruger]

 

[Jianwen Luo]

 

[abdelali belabassi]

i'm etudiant ,i need it to complet my thesis , thanks

[Feng Changjian]

I have like this toolbox because it can help me to complete my school work.

[sina niki]

Dear Sir,
Problem:
 We want simulate by matlab "Passive Radar localization by time difference of arrival esimator".
For this purpose we have two site on the ground and we want by two widwband receivers estimate two same received pulse at separate station.
could you help me to send m.files for this purpose.
thanks your helps.
Regards.

[Jim Green]

I need much higher order spectral analysis toolbox

[Kianoush Nazarpour]

 

[Kim Gunnium]

I need HOSA for learning.

[Nej Nej]

Kindly send me a HOSA tool box for study purpose

[Wei Zhang]

why this toolbox is not concluded in Matlab 7 ?

[Chuanshuang Hu]

It looks great. I need HOSA. Also, could you tell me whether it is possible to compute Auto-and cross-spectra or not using HOSA?

[hawk wei]

Anyway but can you share it ?
It's really great

[farzaneh shayegh]

 

[Marcin J...]

I would be thankfull if you can send me more informations about high order moments. Thank for help.

[ZHONGCHAO HUANG]

It is really good and very helpful to me

[chen xushuang]

very good

[xjtuljp Lin]

 

[xunkai wei]

great

[Bruce Lena]

great tools for me

[Neil L Scott]

Demo fails.
Crashes with GUI callback issues.
Uppercase file names???

[xu guisen]

how to download? who teach me!

[essam anwer]

i can not run this toolbox how can download and who can help me

[Jon-Fredrik Hopperstad]

Excellent toolbox with very good documentation.

[Ghassan Maalouli]

Excellent toolbox. Very good documentation.
Brilliant author.

[meng yusong]

 

[Jon Hobbs]

Great Box

[Sivaganesh Mullapudi]

 

[Mohsen Afrasiabi]

 

[ali dashti]

needed box

[M Nage]

not see yet

[Scott Miller]

How do you open this thing? I have tried to extract it and nothing happens. It seems corrupted.

to reply take out chinese food

[Frederico Melo]

 

[liu xiangle]

the help system is not covinent.

[Vishnuvenkatesh Dhage]

 

[Florencio Calle]

 

[Chang David]

I want to know if the software is only compliant with Matlab 6.0(R12 version)

[song jianjiang]

good article

[yhaofly yhaofly]

your work is execellent,thank you very much,
I will support you for ever.

[ahmed badawy]

pls I hope to make good relation in working
in HOSA and Signal Separation Using HOSA

[Ali Ozgur Argunsah]

hi,
?s there any linux matlab version of this toolbox?

[navya showry]

I have downloaded this file, but .exe file is missing from the extracted one.. plz advice how to proceed further or email me the latest updated one.

[Morteza Dardel]

I want some more description about “nlpow” command (estimation with volterra serises). If this command can be used for more than one input and output cases?

[guang yang]

it's good

[Alhade Mohamed]

hi

[bahman abc]

thx!

[as as]

 

[Vinayak swarnkar]

I downloaded it..but facing problem in running it. Even 'hosademo.m' dosent work...it gives error saying 'Undefined function or method 'choices' for input arguments of type 'char'.'. Can some please help in this.
Thanks you

[Kacimi Abderrahmane]

 

[jason yun]

 

[Guy Incognito]

 

[Alex Yakubovich]

Does anybody what the frequency units are in BICOHER and BISPECD? Some of the examples in the tutorial suggest it is in Hz, but that doesn't make sense given the plot I get with my signal.

[Femi Osunnuyi]

I've had some problems using the 'bicoher' function. With most industrial data sets, i obtain bicoherence values greater than 1, which in theory is not meant to be possible.
Any ideas?
Also is 'segasamp' meant to be square root of no. of data set (which must be even powers of 2 [4,16,64,256,1024], right?)?
  

[Yue ZHOU]

To identify whether one time serial is norm-distribution

[elahe alavi]

 

[Bessie Bao]

EEG

[Edwin Wu]

Does anybody have the experience to use TDE.M? I encounter some problem when I use it as a way to estimate the time delay of 2 channels of signal. I would be so appreciated for your helps.My msn is [e-mail],it is also my E-mail.I would like to make friends with anyone of you.join me , share our knowledge and be brilliant together.

[Gabriel Akira Schreiber]

This toolbox is greatly appreciated. Is this still developed/maintained?

[Gabriel Akira Schreiber]

But, however, lowercase filenames would be a great improvement.

[Rakesh]

I have tried WIG2.m function on a times series on 28800 samples (288 sec signal sampled at 100Hz) but function doesnt seem to handle this big signal. However, I have also tried it on chunks of this signal but it takes forever to complete. Is it my computer or the program itself? any help?

[David Gagnon]

By the way, "segsamp" (actually implemented as "nsamp") is the number of samples per segment, that is, the number of points in each chunk of data that you're throwing into the bicoherence. You can use datasets of any length, and the program will typically calculate fft's with a length that is a power of two and is greater than or equal to "segsamp." It does this by adding zeros to the data record. However, you can specify "nfft" to tell the program how many points you want the fft to use. If you set "segsamp" and "nfft" to be equal, then you can avoid zero-padding. Powers of two are calculated faster, but MATLAB can handle fft's of any length.

As for the frequency axis, "waxis" is based on the number of points that are used in the fft. As a result, "0.5" on the axis corresponds to the highest frequency that can be resolved with that number of points, also known as the Nyquist frequency. Exactly how you want to scale the axis will depend on the nature of your data.

On an unrelated note, the demo gives me errors, too.

[David Gagnon]

I meant to submit this post first...
(Both posts refer to BICOHER.m)

Beware that there are multiple ways to normalize the bispectrum, and they are all referred to as the "bicoherence." The formula that this m-file uses is NOT the one that bounds the bicoherence between 0 and 1. If you are looking for that one, check out Kim and Powers' 1979 paper "Digital Bispectral Analysis and Its Applications to Nonlinear Wave Interactions." They provide an explanation of how to code that version, and it has worked reasonably well for me.

[Paul]

The installation seems to have worked incompletely. I get the following error:
>> HOSADEMO
??? Undefined function or method 'choices' for input arguments of type 'char'.
Error in ==> HOSADEMO at 68
choices('HoSatDemo', 'HOSA Demo', labels, callbacks, interruptible);
My Matlab cannot find the command/function "choices".
I have MATLAB Version 7.8.0.347 (R2009a)
Signal Processing Toolbox Version 6.11 (R2009a)
Can you please help me solve this?
Thanks,
Paul Axon, Sydney, Australia
[e-mail]

[Rae]

Is there a simple way to edit bicoher to normalize it such that it's giving the bounded (0-1) version rather than the current skewness version?

I can't access the Kim and Powers paper and if others have sucessfully done this, I'd love to know how!

[zamrud fathima]

The installation seems to have worked incompletely. I get the following error:
>> HOSADEMO
??? Undefined function or method 'choices' for input arguments of type 'char'.
Error in ==> HOSADEMO
choices('HoSatDemo', 'HOSA Demo', labels, callbacks, interruptible);
My Matlab cannot find the command/function "choices".
Can you please help me solve this?
Thanks.
please send reply to my e-mail