Click here to read the shootout rules

This year’s shootout is quite different from previous competitions. For the first time, a common set of wheat data from three different NIR instrument manufacturers will be analyzed.  The objective is to evaluate the data with the goal of preprocessing the datasets to match the spectra from all three instrument manufacturers prior to developing a regression model for protein that results in equivalent results among the models as measured by the reproducibility.

We would like to thank the U.S. Department of Agriculture’s Grain Inspection, Packers and Stockyards Administration and the instrument manufacturers that participated in the study for providing the data and Dr. Charles Hurburgh, Iowa State University, for facilitating the process.

The most important task will be to determine a method of pre-processing the data to minimize the spectral differences among instrument models so that a single calibration will yield both excellent accuracy and excellent reproducibility among "unknown" instruments of each manufacturer without secondary standardization. The complete set of rules is available here.

Shootout data files:

Manufacturer A (Calibration / Test / Validation)
Manufacturer B (Calibration / Test / Validation)
Manufacturer C (Test / Validation)