How to identify peaks in a chart?

Question

1.00/5 (1 vote)

See more:

Hello
I have a program in C# that takes two vectors X (t) and Y (mV), and plots using MSChart Visual Studio 2010. The plotted graph shows several peaks, would automatically identify the following graph:

-> Beginning and end of each peak;
-> Identify on the program the number of peaks;
-> Identify the value of each peak and at what time it occurs.

Thank you for your attention.

Posted 12-Mar-14 11:20am

jp.51@hotmail.com

Add a Solution

Comments

Sergey Alexandrovich Kryukov 12-Mar-14 17:51pm

Is anything known about the mathematical model of the sampled function? any theory? Because, in its general form, the problem is pretty difficult.
—SA

Matt T Heffron 12-Mar-14 18:08pm

I'd suggest changing your display name NOT to be your email address UNLESS you really like SPAM!
The email address harvesters look EVERWHERE!
When people reply to your question, you'll get an email.

[no name] 12-Mar-14 19:20pm

Do you know any maths? I am serious because the answer really involves maths. You haven't said what it is you are plotting with these 2 (technically) scalars. Is it a continuous function. eg are you reading the output of some device which produces a continuous voltage and sampling at a fixed interval. This should all be in your question.

BillWoodruff 12-Mar-14 22:48pm

Evaluating the relative complexity, or simplicity, of what's involved in achieving your three goals here requires you to tell us more about your code. Specifically, we need to know if you have access to the "raw data:" if you do, then why can't you analyze the data, and develop the summary information you need ?

Second, we can infer from goal three that there is a time-series involved here. Is time one of the axes of the resulting chart ?

Describe the data in detail, and, perhaps, we can assist.

1 solution

Add a Solution

Add your solution here

Treat my content as plain text, not as HTML

Preview 0

…

Existing Members

Sign in to your account

...or Join us

Download, Vote, Comment, Publish.

Your Email
Password
Forgot your password?

Your Email
This email is in use. Do you need your password?
Optional Password

I have read and agree to the Terms of Service and Privacy Policy
Please subscribe me to the CodeProject newsletters

When answering a question please:

Read the question carefully.
Understand that English isn't everyone's first language so be lenient of bad spelling and grammar.
If a question is poorly phrased then either ask for clarification, ignore it, or edit the question and fix the problem. Insults are not welcome.
Don't tell someone to read the manual. Chances are they have and don't get it. Provide an answer or move on to the next question.

Let's work to help developers, not make them feel stupid.

This content, along with any associated source code and files, is licensed under The Code Project Open License (CPOL)

Sergey Alexandrovich Kryukov · Accepted Answer · 2014-03-12T12:02:00

This is not such a simple problem that it may seem.

First of all, you need to totally isolate the analysis of data, and charts, data presentation. Data presentation should only be used as output, never be a source of data. So, let's forget about charts and focus on data.

I don't know the semantic of your data; it does not really matter. For a minute, let's imaging this is some statistical distribution. Even if it is not, you could possibly use the algorithms developed for statistics. Then the peak could be called a mode of distribution: http://en.wikipedia.org/wiki/Mode_%28statistics%29[^].

Your case would be the case of multi-modal distribution: http://en.wikipedia.org/wiki/Multimodal_distribution[^].

The problem appears to be more complex because you don't provide any assumption on the function describing your sample data. In general case, the problem is quite difficult. Please see:
Please see:
http://www.itwm.fraunhofer.de/fileadmin/ITWM-Media/Abteilungen/BV/Pdf/wirjadi08-branch-and-bound.pdf[^],
http://www.stat.tamu.edu/~jchown/NKCreport.pdf[^],
http://books.google.com/books?id=QKxmYCgHn20C&pg=PA58&lpg=PA58&dq=algorithm+find+modes+of+%22multimodal+distribution%22&source=bl&ots=8MVmMHxT-a&sig=avQVC8T-tRwHS3YS6zi76rAVa0w&hl=en&sa=X&ei=X9YgU5n0Nabx0gGhrIDgDg&ved=0CEwQ6AEwBQ#v=onepage&q=algorithm%20find%20modes%20of%20%22multimodal%20distribution%22&f=false[^],
http://books.google.com/books?id=uY-Z3vORugwC&pg=PA54&lpg=PA54&dq=algorithm+find+modes+of+%22multimodal+distribution%22&source=bl&ots=QKu8TbTVz6&sig=glZ6BiCjdbXrV_HzVpPDyJL1biw&hl=en&sa=X&ei=X9YgU5n0Nabx0gGhrIDgDg&ved=0CLIBEOgBMBI#v=onepage&q=algorithm%20find%20modes%20of%20%22multimodal%20distribution%22&f=false[^],
http://www.peerevaluation.org/data/bbf94b34eb32268ada57a3be5062fe7d/tbme-SShahid-2026734-proof.pdf[^].

See also this short reviews of available approaches to general pike detection: http://www.ncbi.nlm.nih.gov/pmc/articles/PMC2631518[^].

Probably, the solution for some special cases would be much easier. For example, maybe your pikes are thing and well isolated; then you could think of something pretty simple, perhaps heuristic.

—SA