Spectroscopic Data

The Unscrambler® has been the software package of choice for many years now for spectroscopic applications. Whether you want to develop quantitative prediction models from Near Infrared (NIR) spectra to classification of metabolites using Nuclear Magnetic Resonance (NMR), The Unscrambler® is fully capable of analysing such complex data and provides users with

• A comprehensive range of data import options from various instrument vendors.
• A complete toolkit of spectral transformations, including

PCA for spectral data
• Smoothing and Derivative
• Scatter Correction (SNV, MSC/EMSC, OSC)
• Normalization and Baseline Correction
• Correlation Optimization Warping (COW) for alignment of data.
• Powerful classification and regression modelling tools, including
  • Principal Component Analysis (PCA) and Cluster Analysis for exploratory data mining.
  • Partial Least Squares (PLS) regression for developing predictive models
  • Advanced classification methods, including the new Support Vector Machines (SVM) classification and Linear Discrimination Analysis (LDA).
• The ability to use The Unscrambler® developed models for prediction of new samples, either within The Unscrambler as a predict/classify option or in real time applications, using the Unscrambler Predictor and Unscrambler Classifier in conjunction with vendor instrumentation.

The Unscrambler® provides the necessary diagnostics and modeling tools seldom found in most multivariate analysis packages and provides the perfect complement to any spectroscopic application.

Sensory Data

The application of multivariate data analysis to sensory data (sometimes referred to as sensometrics) is an ideal application of The Unscrambler®. Data imported from Microsoft Excel sheets or generated using the Quali-Sense can be analysed for the detection of consumer segments through to isolating sensory panellists requiring additional training.

The new L-PLS algorithm can also be applied to sensory data, adding a further dimension of external information for better understanding particular markets. Applications of The Unscrambler® to sensory data include:

Eggshell plot

• The use of Design of Experiment (DoE) for formulating new products ready for sensory assessment
• Using Principal Component Analysis (PCA) for detecting poor panel performers, or for isolating potentially new market segments
• Market segmentation using cluster analysis or the new L-PLS algorithm
• Reverse engineer or define new products in multivariate space
• Comparison of different regions for similar products/service using PCA Projection

The Unscrambler® allows you to weight, standardize or even apply custom factors to input data and then provides advanced graphical output, ensuring the best possible chance of finding new segments, or providing deeper insights into your sensory data.