Title:
Oregano and adulterants NIR

Author:
Rosalba Calvini (University of Modena and Reggio Emilia), rosalba.calvini@unimore.it

Reference:
Ferrari, V., Calvini, R., Menozzi, C., Ulrici, A., Bragolusi, M., Piro, R., Tata, A., Suman, M. & Foca, G. (2024). Addressing adulteration challenges of dried oregano leaves by NIR HyperSpectral Imaging. Chemometrics and Intelligent Laboratory Systems, 249, 105133. https://doi.org/10.1016/j.chemolab.2024.105133

Dataset:
The dataset contains 14400 NIR reflectance spectra (980-1660 nm) extracted from NIR hyperspectral images of dried, ground oregano leaves and four adulterants: myrtle, olive, strawberry tree, and sumac leaves.
The dataset comprises:
- 7200 spectra of authentic oregano (=100 spectra x 24 samples x 3 replicate images);
- 7200 spectra of four adulterants (=600 spectra x 4 adulterants x 3 replicate images).

The spectra are divided into two sets:
- Training: 9600 spectra (4800 oregano and 1200 for each adulterant);
- Test: 4800 spectra (2400 oregano and 600 for each adulterant).

The data were used to develop classification models for oregano authentication following both a class modelling and a soft discriminant approach. Details are given in the following article: Ferrari, V., Calvini, R., Menozzi, C., Ulrici, A., Bragolusi, M., Piro, R., Tata, A., Suman, M. & Foca, G. (2024). Addressing adulteration challenges of dried oregano leaves by NIR HyperSpectral Imaging. Chemometrics and Intelligent Laboratory Systems, 249, 105133.

Typology:
Classification

Dimensions:
14400 samples, 137 variables

Missing values:
No

Columns:
1) Img_ID: Identifier of the hyperspectral image from which the spectrum was sampled
2) Replicate: Image replicate number (R1, R2, R3)
3) Info: Type of material (Oregano, Oregano - inflorescences, Myrtle, Olive, Strawberry tree, Sumac)
4) Class: Class set used for classification (Ad: adulterants, Or: oregano)
5) TR/TS: Training set / Test set partition
6-142) Reflectance at 137 wavelengths (980-1660 nm, with 5 nm resolution)
