I’ve been playing with the spectra and trying to optimize our model, when I found an oddity that I could not explain, maybe someone knows what’s happening here, I’m using the “Hard Cheese” data as my example:
I’ve set the spectra to “normalized” and found a wavelength range that I would like to use as a filter (as seen in the online guide), see picture HardCheese1.
When I enter the filter, the whole spectrum changes, especially th Y-axis, and now it doesn’t look like this wavelength range was the best for one differentiation anymore, but I should’ve rather chosen a range from 1000 to 1060 nm. (see picture HardCheese2)
Can someone help me and explain: Why did this happen an how is this affecting my model?
Thanks a lot!
This topic was modified 7 years, 5 months ago by email@example.com. Reason: Was not finished
The pre-processing methods use some calculations which are only based on the selected range (for example, it calculates the average of the spectrum). If you choose different ranges, the average will be changed – hence the Y axis values will be changed accordingly.