Publication detail

DIMENSIONALITY REDUCTION OF MULTI-VARIATE LASER SPECTROSCOPY DATA

POŘÍZKA, P. KLUS, J. PELASCHINI, F. FABRE, C. PROCHAZKA, D. MOTTO-ROS, V. KAISER, J.

Original Title

DIMENSIONALITY REDUCTION OF MULTI-VARIATE LASER SPECTROSCOPY DATA

Type

abstract

Language

English

Original Abstract

State-of-the-art Laser-Induced Breakdown Spectroscopy (LIBS) instruments enable high repetition rate analysis. With this experimental settings, the mapping of sample surfaces provides large data sets. The richness of information is in spectra (objects) as well as wavelengths (variables). Processing such multivariate data is, thus, a challenging task demanding more sophisticated approaches. Utilization of advanced statistical algorithms, referred to as multivariate data analysis algorithms or chemometrics, are of great interest in contemporary data processing [1-3]. Moreover, elemental composition (relation of individual elements) and structural complexity (relation of individual matrices) provides additional valuable information in understanding of the heterogeneity of, e.g., biological and geological samples. In our work, we bring an introduction to the utilization of Principal Component Analysis to processing of LIBS data. Our efforts tackled mainly the dimensionality reduction in both, objects and variables. Such algorithm leads to an increase in the turn-around time of the multivariate data processing and to a reduction of demands on the computing power. We demonstrated our endeavors on the processing of multi-elemental maps of a geological sample. The original size of the map was 2000x1190 pixels and the spectra contained 2048 variables, resulting in a matrix of 2 380 000x2048 data points. We achieved a conversion to 1 % on the side of objects and 0.44 % on the side of variables of the original data matrix. The data processing brought the same results based on original data matrix as well as on data obtained through such significant information conversion.

Keywords

data, LIBS, reduction

Authors

POŘÍZKA, P.; KLUS, J.; PELASCHINI, F.; FABRE, C.; PROCHAZKA, D.; MOTTO-ROS, V.; KAISER, J.

Released

27. 5. 2018

Publisher

Spektroskopická společnost Jana Marka Marci

ISBN

978-80-88195-06-1

Book

Book of abstracts of 16th Czech – Slovak Spectroscopic Conference

Pages from

60

Pages to

60

Pages count

163

URL

http://16cssc2018.spektroskopie.cz/files/CSSC_2018_BOOK_OF_ABSTRACTS_FINAL.pdf

BibTex

@misc{BUT153343,
  author="POŘÍZKA, P. and KLUS, J. and PELASCHINI, F. and FABRE, C. and PROCHAZKA, D. and MOTTO-ROS, V. and KAISER, J.",
  title="DIMENSIONALITY REDUCTION OF MULTI-VARIATE LASER SPECTROSCOPY DATA",
  booktitle="Book of abstracts of 16th Czech – Slovak Spectroscopic Conference",
  year="2018",
  pages="60--60",
  publisher="Spektroskopická společnost Jana Marka Marci",
  isbn="978-80-88195-06-1",
  url="http://16cssc2018.spektroskopie.cz/files/CSSC_2018_BOOK_OF_ABSTRACTS_FINAL.pdf",
  note="abstract"
}