Options

[HELP] - Data validation for FTIR spectroscopy

productsupproductsup Member Posts: 1 Newbie
Hello RapidMiner community,

I'm in the middle of writing my thesis for data analytics for Fourier-transform infrared (FTIR) spectroscopy. Based on the data that I have from the researchers, I can see the similarities of the same object in terms of its own spectrum (See pdf attached file).

My hypothesis is: If we see the similarities in the second experiment towards the first experiment, then we don't need to run the experiment for the same object again for the third time. Right now, the reason why the researchers running it multiple times was that they want to have confidence in their data. But if you see with the naked eyes, the spectrum almost identical between the first and the second one.

My question here how do I validate my hypothesis here? Does someone encounter a similar issue before? Thanks in advance guys!

Answers

  • Options
    Telcontar120Telcontar120 Moderator, RapidMiner Certified Analyst, RapidMiner Certified Expert, Member Posts: 1,635 Unicorn
    There isn't any built-in operator to do this, but it is feasible, I think.  You would need to take the underlying values at specific points in the spectrum (and you could select as many as you want) and then use RapidMiner to compare the similarity between the two using a metric like cross-distances.
    @sgenzer used to work with FT and may have some additional ideas.
    Brian T.
    Lindon Ventures 
    Data Science Consulting from Certified RapidMiner Experts
Sign In or Register to comment.