Choose Language Hide Translation Bar

Using JMP® Partial Least Squares to Model Chemometric (Near-Infrared Spectrometry) Data to Predict Percent Flavor on Coffee


Chris Liebold, PhD, Scientist, The J.M. Smucker Company

During the production of flavored coffee products, it is useful to have a fast, reliable and effective quality control analysis for the amount of liquid flavor added to the coffee. Near-infrared spectrometry offers the ability to quickly analyze products, which is essential for quality control analysis. It generates a unique spectrum that has 1,049 data points that create the unique signature for the sample analyzed. However, the instrument's ability to predict percent flavor is solely based on the calibration model developed for it. The Partial Least Squares platform in JMP was executed to generate a percent flavor model that can take advantage of the speed and unique data generated by a near-infrared spectroscopy instrument.



Can I email you with some questions re your presentation?

If yes, what is your email address?

Thank you,



I need any tutorial for this presentation, can you help me please?

@mca, the only materials for this presentation is the PDF that is attached.


However, the Mastering JMP webcast series covered Partial Least Squares in Using Partial Least Squares - When Ordinary Least Squares Regression Just Won’t Work.




Thank you very much, Jeff...


I watched webcast series. But, i am interesting for use of JMP with spectral data. The presentation above is very useful but need some detail.



Article Labels
Article Tags