Multivariate Analysis

Multivariate Icon

Advanced data processing for characterization of complex sample systems is available in LabSpec 6’s Multivariate Analysis (MVA) module. It includes a number of multivariate (chemometric) methods which are seamlessly powered by Eigenvector Research Inc. a world leader in chemometric and multivariate technology.

Key methods which are included are:

  • Classical Least Squares (CLS) Fitting
  • Principal Components Analysis (PCA)
  • Multivariate Curve Resolution (MCR)
  • Hierarchical Clustering Analysis (HCA)
  • Divisive Clustering Analysis (DCA)
  • Partial Least Squares (PLS)

These methods can be used for many varied applications, but typically offer powerful solutions for more complex datasets where traditional univariate techniques cannot adequately describe the sample variation or composition.

The six available methods can be used for decomposition of samples (e.g., to understand the distribution of components in a Raman map using CLS, PCA and MCR), clustering (e.g., to classify and group related spectra/samples together using HCA and DCA) and quantitative characterization (e.g., to create quantitative calibration models based on spectra of known concentrations, and then apply the model to target data sets to predict concentrations where they are not known, using PLS).

The module is fully integrated within LabSpec 6, allowing users to acquire data, analyse and report from within a single platform.  The intuitive interface requires minimal knowledge of multivariate techniques; commonly used pre-processing functions for the included methods are available as part of the MVA module, whilst LabSpec’s full data processing suite is available for users who wish to customize their analyses.

multivariate methodologies
Figure 1: screen shot showing a Raman hyperspectral map with univariate (cursor) and multivariate (MCR) analysis results. Loading spectra and statistics are displayed in the right hand tab.

The LabSpec 6 MVA module intentionally offers only a small selection of available multivariate methodologies, thus allowing the module to be focussed on the needs of typical users. In cases where additional methods and advanced control is required, HORIBA Scientific recommends Eigenvector Research’s Solo+MIA software, a versatile and leading standalone package for multivariate analysis. LabSpec 6 offers a one click data link, allowing users to transfer data directly from LabSpec 6 into Solo+MIA at the touch of a button.