0

I hope my question is within the topic of this forum.

I would like to test an algorithm to a real application data and publish in a journal/conference. However, I cannot acquired my own data set myself and I am thinking to use one of Matlab's demos data set that is mentioned from a real experiment.

Is it acceptable to use those data set for a conference/journal publication?

Thank you

Tanpa Nama
  • 21
  • 1
  • 2
    Are you asking whether the license to the Matlab software allows you to publish work based on their data set? If so, what does the license say? Or are you asking whether this is valid from a scientific perspective (we can't possibly answer that, since we don't know your research)? – ff524 Aug 26 '15 at 06:44
  • I did not think of your first question, but my question is the second: the validity from a scientific perspective. I am doing system identification research – Tanpa Nama Aug 26 '15 at 06:51
  • 2
    Well, we don't know anything about the research or the dataset, so it's not really possible to judge. The fact that it's a Matlab data set is not relevant; the question is whether it's an appropriate data set for your research, and that is out of scope of this site. – ff524 Aug 26 '15 at 06:52
  • After rethinking my question, I think I have asked an out of scope question. I am sorry – Tanpa Nama Aug 26 '15 at 07:00
  • 1
    If the data is "real" (you mention that it comes from a real experiment) and it's stated where it comes from, you should try to find the original source of the data (research paper, report etc) and refer to that instead of referring to "Matlab demo data". The question about suitability for your application etc that @ff524 mentions is still valid though. – fileunderwater Aug 26 '15 at 08:07
  • Using widely available data sets is certainly useful, so others can check your results or compare other approaches. – vonbrand Aug 26 '15 at 21:05

1 Answers1

2

As an alternate suggestion, you might try calling people who either do similar research, or are involved in the application the research covers, and get a copy of their data. I've never done it at a publishing level, and I don't know what kind of field you're in, but I have had success with this for some minor projects I've done.

Generally, it's not in the exact format you want, but real data is better than suspicious data. And reformatting is a lot easier than collecting tons of data.

MichaelS
  • 229
  • 1
  • 3