Impact of data reduction on multivariate classification models built on spectral data from bio-samples

Abstract

Multivariate data analysis methods have been used to evaluate single shot spectral data, obtained by laser induced breakdown spectroscopy (LIBS), from ten different biological samples (simulants and possible interferents in Biological Warfare Agent (BWA) detection applications). Spectral data as echellograms (2D CCD images) and extracted 1D spectra were used and the classification performance was studied as the number of input variables was altered. Principal component analysis (PCA) indicated a possibility to separate the samples due to spectral differences, and partial least squares discriminant analysis (PLS-DA) was applied to study the predictability in more detail. For full resolution 1D spectra, a normalization of the data mainly resulted in visual effects in the PCA score-plots without significant effect in predictability by the PLS-DA models, however, normalization improved the predictability if the amount of variables were heavily reduced. A quite strong data (variable) reduction could be performed on both the 1D and 2D data without losing significant predictability. Using similar amounts of variables, the prediction models performed better using the echellograms directly compared to the extracted 1D spectra. The problem of spectral data shift (relative 'database' spectra) was also investigated, where already small shifts cause the models to fail. However, after a selection of important variables and allowing certain regions for these variables, the impact of shift on predictability could be reduced.

Femtosecond Single-Pulse and Orthogonal Double-Pulse Laser-Induced Breakdown Spectroscopy (LIBS): Femtogram Mass Detection and Chemical Imaging with Micrometer Spatial Resolution

“…Furthermore, binning allows for faster image acquisition as the number of readouts is reduced by n 2 . [49][50][51][52][53]…”

Multivariate classification of echellograms: a new perspective in Laser-Induced Breakdown Spectroscopy analysis

“…The idea of direct echellogram utilization was originally suggested by Larson et al . 27 They proved the superiority of echellograms, especially after cropping and binning, over the conventional utilization of 1D spectra for classification purposes. A significant reduction in the data readout from the ICCD and transfer times was yielded by binning the pixels and/or cropping the image, which could induce a higher sampling rate.…”

“…Larsson et al . 27 compared the performance of a Partial Least Squares Discriminant Analysis (PLS-DA) algorithm where either 1D spectra or raw echellograms were fed to the algorithm as the input variables. Their aim was to show the impact of variable reduction on the discrimination capability.…”

Hadamard patterned masks for large-aperture UV Raman spectroscopy in low volatile CWA detection applications

“…Examples of such substances could be highly toxic material (e.g., Chemical Warfare Agents, CWA or Pharmaceutical Based Agents, PBA) or different types of chemicals related to production and handling of Home Made Explosives, HME. We have earlier explored optical spectroscopy for the detection of hazard materials, e.g., CWA [1], Biological Warfare Agents (BWA) [2], and energetic materials [3], including calibrated measurements resulting in knowledge of fundamental parameters such as molecular Raman cross-sections in the UV [4]. In addition, feasibility of fluorescence and plasma emission spectroscopy for bioaerosol detection [5,6], as well as hyperspectral imaging via Compressed Sensing and Coded Apertures for possible applications in future detection systems [7] have been investigated.…”