I have been working on creating a multivariate predictive model (most likely PLS) using FTIR spectra to predict the amount of a component in complex mixtures. In order to calibrate the model I have considered two pathways - creating synthetic mixtures of the component and confounders or using an established technique to give me the true value of the component in real mixtures. Both pathways have big limitations. I am now considering spiking real mixtures with several known (by weight) amount of the component. How can I implement this in a X-Y chemometric model ?
Is there a reason you're setting this up as PLS model instead of a traditional concentration curve? What are your variables? How many variables do you have?
Let's consider simplistic case with 5 vars: A, B, C, D, E, and that var A is what you're trying to predict. If the other vars don't correlate with A, then you could reduce this to a calibration curve method. If, for example, var C is producing or catalyzing the production of A (i.e., A and C are heavily correlated), simply adding A at various amounts of a synthetic mixture may not produce an accurate model, because it's difficult to tell the difference between what was just added, and what's being produced by correlating vars.
Would it be possible to collect data from both synthetic mixtures and real samples? Running these against each other could have some valuable information, particularly if they were to separate in space. If they don't separate in space, then your synthetic mixtures may be a good representation of real samples.
Kirk,
Thanks for the comment!
I want to use PLS or some other multivariate chemometric technique for their ability of using tons of variables for the predictive model. In the case of FTIR, generally the integration of one spectral region is used in a traditional concentration curve. In the case of PLS, I can use the entire spectrum (or a good portion of it) and select the important LV.
Probably for you variables are the components in the mixtures. In my case, there are generally between 5-10 components in the mixture. I am interested in the quantification one of them only. The other 4-9 components create(correlate) with the signal of my target in the entire spectrum. The issue with synthetic mixtures is the poor quality of the standard materials (with the exception of my target) and limitation of the techniques to determine the composition of the standard materials. The issue of understanding exactly what is in the mixture is why I cannot use simply real samples.
The standard addition method could be the best way ....but....I have not figured out how to do it chemometrically.
Does it make it sense?
- Badges
- Science topic
- Similar topics
- Biological Science
- Zoology
- Ornithology