Theses and Dissertations
Issuing Body
Mississippi State University
Advisor
Philip, Thomas
Committee Member
Singh, J. P.
Committee Member
Chu, Yul
Date of Degree
8-2-2003
Document Type
Graduate Thesis - Open Access
Major
Computer Engineering
Degree Name
Master of Science
College
James Worth Bagley College of Engineering
Department
Department of Electrical and Computer Engineering
Abstract
Laser-induced breakdown spectroscopy (LIBS) is an advanced data analysis technique for spectral analysis based on the direct measurement of the spectrum of optical emission from a laser-induced plasma. Assignment of different atomic and ionic lines, which are signatures of a particular element, is the basis of a qualitative identification of the species present in plasma. The relative intensities of these atomic and ionic lines can be used for the quantitative determination of the corresponding elements present in different samples. Calibration curve based on absolute intensity is the statistical method of determining concentrations of elements in different samples. Since we need an exact knowledge of the sample composition to build the proper calibration curve, this method has some limitations in the case of samples of unknown composition. The current research is to investigate the usefulness of ANN for the determination of the element concentrations from spectral data. From the study it is shown that neural networks predict elemental concentrations that are at least as good as the results obtained from traditional analysis. Also by automating the analysis process, we have achieved a vast saving in the time required for the data analysis.
URI
https://hdl.handle.net/11668/19793
Recommended Citation
Inakollu, Prasanthi, "A Study of the Effectiveness of Neural Networks for Elemental Concentration from Libs Spectra" (2003). Theses and Dissertations. 293.
https://scholarsjunction.msstate.edu/td/293