puredata: Datasets for spectral unmixing examples
In svkucheryavski/supure: Spectral Unmixing with Pure Variables

Description Usage Format Details References

puredata contains four different spectral datasets used in the chapter and also in help examples of this package.

1	data(puredata)

There are four datasets:

`carbs`	generated Raman spectra of ribose, fructose and lactose mixtures.
`simdata`	generated UV/Vis spectra of polyaromatic hydrocarbons mixtures.
`mir`	Mid infrared spectra of ethanol and glucose.
`hsi`	Raman hyperspectral image of oil-in-water emulsion.

Each dataset is a list with following fields:

`D`	data matrix with spectra of mixtures.
`S`	optional, matrix with spectra of pure components.
`C`	optional, matrix with concentrations of pure components.
`wavelength`	wavelength or wavenumbers.

The first dataset, carbs, consists of Raman spectra of fructose, lactose and ribose generated from individual spectra with relative concentrations followed a simplex lattice design with four levels. The individual spectra were downloaded from publicly available SPECARB library [1]. The spectra have clear pure variables for each component, e.g. 818 cm-1 for lactose, 357 cm-1 for fructose and 542 cm-1 for ribose).

The second dataset, simdata, consists of UV/Vis spectra of three polyaromatic carbohydrates and their mixtures. Values for concentration of the components in the mixtures were generated randomly, using uniform distribution with limits from 0 to 1 (Mole). The pure spectra can be resolved using second derivative and following set of pure variables: 267 nm, 303 nm and 311 nm.

The third dataset, mir contains IR spectra of binary aqueous ethanol-glucose solutions. The spectra were acquired with a Matrix-MF FT-IR spectrophotometer with a diamond crystal probe produced by Art Photonics GmbH in a range 680...2900 cm-1 [2]. There are no pure spectra or concentrations available for this dataset.

The fourth dataset, hsi, is a 60x60 pixel Raman hyperspectral image of an oil-in-water emulsion [3]. The depicted area consists of an aqueous phase (with mostly glycerol and in a smaller concentration polawax), two droplet phases (one is polawax and another with parabens, polawax and some panthenol) and a structural phase with a crystalline polawax. The matrix with data, D, in this case has two additional attributes, height and weight for height and width of the image in pixels, which allows to recognize this data by the methods implemented in the package as an image.

1. http://www.models.life.ku.dk/~specarb/specarb.html

2. A. Bogomolov, et al., Sensors Actuators B Chem. (2015).

3. J.J. Andrew, M. a. Browne, I.E. Clark, T.M. Hancewicz, A.J. Millichope, Appl. Spectrosc. 52 (1998)

svkucheryavski/supure documentation built on May 30, 2019, 9:32 p.m.