近红外光谱信息网,近红外光谱,近红外,NIR,近红外光谱专业委员会

The Calibration Wizard allows fast and easy development and validation of calibrations without any chemometric knowledge. The Calibration Wizard automatically finds the optimal parameters for your calibration from a large number of possibilities. The quality of the individual Wizard calibrations is determined via the NIRCal Q-Value (Calibration Quality Value). The Calibration Wizard and the Q-Value are the result of our vast, accumulated chemometric expertise. The patented Calibration Wizard leads even inexperienced users quickly and completely to powerful and stable calibrations - without requiring any in-depth chemometric understanding.

For validation of the model, you can define a separate validation set. In addition, you also have the option of Cross Validation. Various statistical parameters (SEC, SEP, BIAS, PRESS, r2, … ) assist you in validating calibrations. Statistical tests such as Durbin-Watson or t-Test help you interpret data with a higher degree of reliability. The Durbin- Watson Test checks quantitative methods for linearity. The t-Test checks the equivalence of the validation sample results versus the reference values.

Various reports are available for external validation for quantitative and qualitative calibrations. Accurate tests and optimizations are possible with the aid of the listing of non-identification, misidentification, and additional valuable data.

• Multivariate quantitative and qualitative (classification, identification) calibration methods
  PLS, PCR, MLR and CLUSTER, SIMCA.
  Multivariate data analysis with PCA.
• Large selection of 45 data pre-processing algorithms (pretreatments) that can be used in any sequence, e.g. normalization, SNV, derivatives (1st, 2nd, 3rd), smoothing, transformations, offset correction, MSC, custom linear filtering
• Arbitrary free wavelength (variable) selection possible (pretreatments depends on wavelength selection)
• Validation Set or Cross Validation with different grouping algorithms and jack-knifing
• Detailed Calibration Protocol including calibration version, all parameter settings, prediction results, used spectra
• Prediction protocols of selected external spectra to compare calibration performance and robustness
• Outlier Detection in modelling and prediction (Property, Residual, Leverage, Score)

Data Visualization is essential for data exploration, outlier detection and data cleaning.
In NIRCal highly interactive graphical visualization of spectral data and all type of chemometric parameters and results can be combined in different plot types

• line plot
• parallel coordinates plot
• scatter plot in 1D, 2D and 3D
• surface plot in 2D and 3D
• data table 

and interaction with

• selecting
• zooming
• panning
• scrolling
• rotating
• resizing
• coloring
• visibility
• data info tooltips

Easy selection of data sets (spectra, properties, wavelength, principal components) in graphics and tables. 

• Data link within all plots as well as tables.
• Immediately switching between wavenumber [1/cm] and wavelength [nm].
• Overview plots
• Dependency plots (data versus time, creator, instrument, property)
• Fast calculation speed
• Graphics for large amount of data

Integrated BASIC macro programming environment to customize NIRCal directly or externally via COM interface to access the predictor, custom import/export, data conversion, data selection, calibration optimization.