Code covered by the BSD License  

Highlights from
Hunt For Local Maxima, Minima, Plateau

5.0

5.0 | 2 ratings Rate this file 15 Downloads (last 30 days) File Size: 217 KB File ID: #29432
image thumbnail

Hunt For Local Maxima, Minima, Plateau

by

 

17 Nov 2010 (Updated )

Illustrates identification of local maxima, minima or plateau, for exit criteria for long processes.

| Watch this File

File Information
Description

======================================
main executing reference usage: usage_proclivityStateTracking.m
* Each trend data point are to be of large time interval
----------------------------------------------------------------------------
The objective is to illustrate identification of a local maxima, minima or plateau.
usage_proclivityStateTracking.m demonstrates a situation where the output of each point may have an interval of hours.

A braking condition may be posited to halt at certain conditions (braking or exit criteria), for example,
after a number of maxima, minima or plateau.

Given a program encapsulated by benchmarkSampleFunction.m (short naive process is used in place of long processes for illustrative purposes, you wish to replace with your own processes),
the demo illutrates the utility required for long processes where each round of output takes a long time.
User may wish to halt in condition of the detection of the next maxima. minima, plateau start or plateau end.
Further conditions may start N maxima, N minima, N plateau start/ end from now.

2nd demo (trivial): usageMain.m
* All trend data points are available
----------------------------------------------------------------------------
Illustrates the mechanisms of the signal feature (maxima, minima or plateau) hunting.

* Caveat :
This build is not a perfect build, it serves as an illustration on which a framework may be built upon with the feature hunter functions, such as:
1. determineLocationsOfMaximaMinimaPlateauFromProclivityFeature.m : only when the a range of data points with it's signal trend/ profiles are available.
2. determineProclivity.m : given 2 data nodes or points, determine signal trend, ie. {'plateau', 'ascent', 'descent'}

Limitations:
The builds work for most fluctuating data signals. However, there are some signals which may cause some exceptions.
The author appreciates suggestions and errata. Please do not hesitate to send suggestions and send feedback for improvement for the framework construction to the email provided.

If the demo has more elegant presentation, please do not hesitate to suggest and send feedback to author.
Email: promethevx@yahoo.com.

Thank you.

Regards,
Michael Chan JT

MATLAB release MATLAB 7.10 (R2010a)
Tags for This File   Please login to tag files.
Please login to add a comment or rating.
Comments and Ratings (2)
29 Jan 2011 Sean

A useful feature for biomedical utility. Could you provide a C code version for this as well? I have just sent you an email. Thank you.

10 Dec 2010 Yen Hanning

Nice. I can incorporate this to my work.

Updates
19 Feb 2012

add tags

Contact us