We make NIR Chemometrics easy

Hi, we’re CalibrationModel. Our aim is to transform your NIR data to superior calibration models. We do this by using knowledge driven software applying good practices and rules from literature, publications, regulatory guidelines and more. Our service is used by NIR specialists to deliver a valuable model for their NIR analysis measurements. With CalibrationModel services, NIR specialists can find out how their NIR Data can be robust and optimally modeled by which data preprocessing and wavelength selection, etc. You can implement CalibrationModel in a matter of minutes using our contact form and send your NIR data to receive optimized model settings as a blueprint.
NIR specialists (Spectroscopist, Chemometricians) love perfect models. They’re curious about how to improve their models even further, because all NIR models need continuous maintenance and updates.
Using CalibrationModel services, NIR Specialists can deliver real value to their measurement results through powerful model optimization capabilities.
CalibrationModel We make NIR Chemometrics easy. Near-Infrared Data Modeling Calibration Service

Procedures for NIR calibration – Creation of NIRS spectroscopy calibration curves

Do you know the effect that you prefer to try out their favorite data pretreatments in combination and often try the same wavelength selections based spectra of the visualized?

You try as six to ten combinations until one of them selects his favorite calibration model, to then continue to optimize. Since then suddenly fall to outliers, because it goes in depth, so is familiar with the data, we know now the spectra of numbers of outliers and is familiar with the extreme values.

Now, the focus is on the major components (principal components, Latent Variables, factors) and makes sure not to over-fit and under-fit not to. The whole takes a few hours and finally one is content with the model found.

So what would happen if you all in the beginning tried variants found outliers removed and re-evaluated and compared? The results would be better than that of the previous model choice? One does not try out? Because it is cumbersome and takes hours again?

We have developed a software which simplifies this so that also the number of model variations can be increased as desired. The variants generation is automated with an intelligent control system, as well as the optimization and comparing the models and finally the final selection of the best calibration model.

Our software includes all the usual known data pretreatment methods (data pre-processing) and can combine them useful. Since many Preteatments are directly dependent on the wavelength selection, such as the normalization the determined within a wavelength range of the scaling factors to normalize the spectra so that pretreatments with the wavelength ranges may be combined. So a variety of settings sensible model comes together that are all calculated and optimized. For the automatic selection of the relevant wavelength ranges, different methods are used, which are based on the spectral intensities. Thus, for example, regions with total absorption is not used, and often interfering water bands removed or retained.

Over all the calculated model variations as a summary outlier analysis can be made. Are there any new outliers (hidden outlier) discovered, all previous models can be automatically recalculated, optimized and compared without these outliers.

From this great number of calculated models with the statistical quality reviews (prediction performance) the optimum calibration can now be selected. For this purpose, not simply sorting by the prediction error (prediction error, SEP RMSEP) or the coefficient of determination (coefficient of determination r2), but by several statistical and test values are used jointly toward the final assessment of optimal calibration.

Thus we have created a platform that allows the highly automated work what a man can never do with a commercial software.

We therefore offer the largest number of matched to your application problem modeling calculations and choose the best calibration for you!

This means that our results are faster, more accurate, robust and objective basis (person independent) and quite easy for you to apply.

You have the full control of the models supplied by us, because we provide a clearly structured and detailed blueprint of the complete calibration, with all settings and parameters, with all necessary statistical characteristics and graphics.

Using this blueprint, you can adjust the quantitative calibration model itself in the software you use, understand and compare. You have everything under control form model creation, model validation and model refinement.

Your privacy is very important to us. The NIR data that you briefly provide us for the custom calibration development will remain of course your property. Your NIR data will be deleted after the job with us.

Interested, then do not hesitate to contact us.

NIR Spectroscopy Calibration Report for quantitative predictive models

When you send your quantitative NIR spectra data to our NIR Calibration Model Service, you get a detailed calibration report (calibration protocol) of the found optimal calibration settings, so you are able to see all insights and easily re-build the model in your NIR/Chemometric software.

Here is a part of our calibration report, that exactly describes the data used in the calibration set (CSet), the validation set (VSet) and the test set (TSet). The numbers are the number ids of the spectra in your delivered NIR data file.


The calibration method settings and parameters are
Waveselection : the variable selection or wavenumber selection or wavelength selection
Pretreatments : the spectral data pre-processing
PCs : the number of  Principal Components (PC) or Latent Variables (LV)
Method : the modeling method algorithm used, e.g. PLS

Then the statistical analysis of the PLS model by the different sets (CSet, VSet, Tset).

Calibration Report

Statistical analysis of calibration, validation and test results : 1 Name, 2 Unit, 3 N : number of spectra, 4 N : number of samples, 5 Average spectra count per sample, 6 Reference values, 7 Min, 8 Mean, 9 Median, 10 Max, 11 Standard deviation, 12 Skewness : left (-) or right (+) lack of symmetry, 13 Kurtosis : flat (-) or peaked (+) shape, 14 Model statistics, 15 RPD, 16 R², 17 RMSEC, RMSEP, RMSET : root mean square of prediction errors, 18 SEC, SEP, SET : standard error (bias corrected), 19 Bias, 20 Skewness of prediction errors, 21 Kurtosis of prediction errors, 22 Intercept, 23 Slope, 24 Intercept (reverse), 25 Slope (reverse), 26 Sample Prediction Repeatability Error, 27 Sample Prediction Repeatability Error (of Missing data MSet)

This shows how we deliver the optimal settings. With the statistical values, the NIR model predicted values of all spectra and additional plots you are able to compare with your re-built model to verify that the models perform nearly equally.

How to develop near-infrared spectroscopy calibrations in the 21st Century?


The Problem

Calibration modeling is a complex and very important part of NIR spectroscopy, especially for quantitative analysis. If the model is badly designed the best instrument precision and highest data quality does not help getting good and robust measurement results. And NIR Spectroscopy requires periodically recalibration and validation.


How are NIR models built today?

In a typical usage in industry, a single person is responsible to develop the models (see survey). He or she uses a Chemometric software that has a click-and-wait working process to adjust all the possible settings for the used algorithms in dialogs and wait for calculations and graphics and then to think about the next modeling steps and the time is limited to do so. Do we expect to find the best use-able or optimal model that way? How to develop near-infrared spectroscopy calibrations in the 21st Century?


Our Solution

Why not put all the knowledge a good model builder is using into software and let the machines do the possibilities of calculations and presenting the result? Designing the software that way, that the domain knowledge is built-in, not just only the algorithms for machine learning and make it possible to scale the calculations to multi-core computers and up to cloud servers. Extend the Chemometric Software with the Domain Knowledge and make as much computer power available as needed.

As it was since the beginning

User  → Chemometric Software → one Computer → some results to choose from

==> User’s time needed to click-and-wait for creating results

Our Solution

User → (Domain Knowledge → automatized Chemometric Software) → many Computers → the best models

==> User’s time used to study the best models and reasoning about his product / process

Note that the “Domain Knowledge” here does perfectly support the User’s product and process knowledge to get the things done right and efficient.


Scaling at three layers

  • Knowledge : use the domain knowledge to drive the Chemometric Software
  • Chemometric Software : support many machine learning algorithms and data pre-processings and make it automatic
  • Computer : support multi-core calculations and scale it to the cloud

The hard part in doing this, is of course the aggregation of the needed domain knowledge and transform it into software. The Domain Knowledge for building Chemometric NIR Spectroscopic models is well known and it’s huge and spreads multiple disciplines. Knowledge-driven software for computing helps to find the gold needle in the haystacks. It’s all about scaling that makes it possible. See Proof of Concept.


New possibilities

  • NIR users can get help working more efficient and getting better models.
  • New types of applications for NIR can be discovered.
  • Evaluation of NIR Applications to replace conventional analytical methods.
  • Hopeless calibrations development efforts can be re-started.
  • Higher model accuracy and robustness can be delivered.
  • Automate the experimental data part of your application study.
  • Person independent optimization will show new solutions, because it’s not limited by a single mindset => combining all the aggregated knowledge and its combinations.
  • Software independent optimization will show new solutions, because none of vendor specific limitations and missing algorithms are present => combining all open available algorithms and there permutations.
  • Computing service is included.

Contact us for trial

Your NIR data is modeled by thousands of different useful calibration models and you get the best of them! That was not possible before in such a easy and fast way! See How it works

Summary of the NIR Chemometric survey polls

Summary of the NIR Chemometric survey polls (as of end of Sept. 2013)

The interesting finding is that most of the answers fit the following pattern. The most companies that use NIR have one NIR Instrument and only one employee that is able to develop NIR calibrations. For that the most common off-the-shelf chemometrics program is used and spent 2 hours or over a month and therefore gets no calibration training about the complex topics like Chemometrics and NIR Spectroscopy or only once (introduction). The calibration maintenance ranges from never to 3 times a year. Interestingly, there was no one who uses portable NIR instruments. We continue our surveys, for the discovery of new trends. Conclusion Seeing this picture, we think that there is huge potential to improve the calibrations. Advanced knowledge can help individuals to build the calibrations with best practices and improve their models accuracy and reliability. Once the decision and investment in NIR technology is done, you should get the best out of your data, because this extra NIR performance can be given by calibration optimization. We offer this as an easy to use and independent service.

What is a NIR calibration used for?

NIR calibrations are used for NIR contents analysis as a productive analytical method. That is a two step procedure.
  1. A NIR analyzer does a non-destructive optical scan of a sample that yields a measured spectrum in seconds.
  2. A NIR calibration model can quantitatively predict (analyze, determine, estimate) multiple constituents, ingredients, contents, analytes, assay, API and other parameters and attributes (chemical, physical, biological, biochemical, sensory) summarized as properties, out of a single spectrum in milli seconds.
The NIR analysis is a very fast non-destructive analysis method that can replace or backup slower methods like wet chemical analysis, chemistry laboratory, sensory panels or rheology (viscosity). Or a NIR calibration can open the door to new possibilities of analytics, quality assurance and process control, by developing calibration models for parameters that seems to be impossible, because they are based on human knowledge, empirical values or sensory like taste value. If you have an NIR instrument, you can measure your samples systematically and thus develop your own calibration models.

What are pre-developed NIR pre-calibrations?

There are a lot of terms that means the same, pre-calibration or NIR starter calibration or pre-built calibration or pre-installed calibration orcalibration package or pre-developed calibrations or pre-calibrated NIR or global calibrations or nir global calibration package or factory calibrations or universal near-infrared (NIR) calibrations or local calibrations or ready-to-use NIR calibrations or off-the-shelf calibrations or factory-calibrated or pre calculated model or start-up calibrations or calibration equations or prefabricated nir calibrations or calibration library or mathematical model. That are Calibration models that are prepared and developed by a calibration specialist. They have collected a lot of samples over years and measured them with NIR and analyced it with reference methods. The NIR spectra are then calibrated against the reference values. This is called a NIR calibration or calibration model or sometimes calibration curve or calibration equation. Normally a precalibration is delivered as a file that is compatible to the used NIR analysis software. Such a calibration file does not contain the spectra nor the reference values.

So how can that work?

The only thing that is in the file is a description what it is for (e.g. protein in feed) and the chemometric model that is represented and stored as list of vectors and matrices. You can’t visualize them, it’s a black-box file. You have no insight of how the calibration is done, how are the settings, how is the prediction performance. You can not extend the calibration with your data to adjust it to your purpose or specialty. Most often the pre calibration files are protected, so you can use it only with a paid license to your software or even to your instrument serials number. These are some (not well known) limitations you will discover if you got one. But such starter calibrations are very useful to have a fast and easy start with a new NIR spectrometer. That’s the main reason why pre-calibrations are available. The second reason is that a collection of spectra can be reused to build such pre calibrations.

Predicting the future?

Are very old spectra useful to predict the future? To adjust a calibration model with newly collected data, the calibrations grows and contains more and more redundancy. That means there are very similar spectra with the same concentration range. So which spectra can be removed to make the calibration better? You maybe never ask this because often you hear, that the more spectra you put into a model the better it will be. Why to remove some spectra?
  • reduce not needed redundancy
  • makes the calibration smaller and less complex
  • makes the calibration better fit to the current situation of now and the near future
  • remove long past seasonal data if you have natural products because nature is changing
  • and of course bad outliers should be removed

Custom NIR calibrations

Build your own calibrations that perfectly fit to your specific sample matrix of your products and your preferred raw materials from your local suppliers. Nature grows differently depending on the geographical region, by seasons and year by year. As you know that NIR-Spectroscopy is not an absolute method, then you have to think about to calibrate these current changing effects into your models. If you own the spectra and the reference values then your are able to build your own calibration models and re-calibrate them when needed. So you have the full control on Calibration updates (also known as moving models).

Conclusion

A NIR-instrument can only measure NIR spectra. So the usefulness of NIR comes in with calibrations. That is very important to know when buying such an instrument. For a fast start you can use pre-built calibrations. Good reliably calibrations are offered from third party to quite high prices that level is similar to a cheaper NIR-Instrument! To continue successfully it is highly recommended to develop your own customized calibration (multivariate calibration model) with your own data from your own products, especially with the use of natural resources. Therefore you need knowledge about chemometrics and multivariate analysis (MVA), spectroscopy and the software used to get the calibration optimized. It is worthwhile to create your own calibrations, because you can calibrate product characteristics that are not covered by the proposed pre-calibrations.

What is JCAMP-DX ?

JCAMP-DX is a Electronic Data Standards for long-term storage and transfer of chemometric information. The standard is development by the International Union of Pure and Applied Chemistry (IUPAC).

JCAMP-DX is an abreviation for the Joint Committee on Atomic and Molecular Physical data – Data eXchange.

It is an human readable file format that is used to store near infrared spectrometry data (and others like Raman, UV, NMR, mass, x-ray, chromatograms, thermograms) and related chemical and physical information and is used since the late 80s.

Almost all NIR-software packages can export the spectra including the reference values as JCAMP-DX. A single file can contain multiple spectra and reference values. A JCAMP file name looks like “sample.dx“, “sample.jdx” or “sample.jcm“.

All data are stored as labeled fields of variable length using printable ASCII characters. Such files can be loaded in an text editor to check the content:


##TITLE= Indene (FILE: AFIR2.DX)
$$ FILE AFIR2.DX ( derived from TFIR2.DX)
$$ ABSORBANCE
$$ FIXED FORM
$$ INCREASING ABSCISSA
$$ RATIONAL ABSCISSA SPACING
##JCAMP-DX= 4.24 $$ Encoded by INTTODX 1.04 (RS McDonald)
##DATA TYPE= INFRARED SPECTRUM
##ORIGIN= JCAMP-DX Test Disk 1.04
R.S.McDonald, 9 Woodside Dr., Burnt Hills, NY 12027, 518-399-5145
##OWNER= Public Domain
##RESOLUTION= 2.0
##DELTAX=1.00000000
##XUNITS= 1/CM
##YUNITS= ABSORBANCE
##XFACTOR= 1.000000000
##YFACTOR= 0.000100000
##FIRSTX= 400.000
##LASTX= 4.000E+03
##NPOINTS= 3601
##FIRSTY= 3.487E-1
##XYDATA= (X++(Y..Y))
400 3487 3355 3264 3198 3153 3143 3182 3298 3520 3845 4262
411 4783 5449 6304 7383 8684 10209 12041 14123 16003 16162 14191
422 11791 9674 7943 6540 5406 4528 3874 3397 3045 2780 2584
433 2446 2354 2290 2246 2212 2187 2165 2135 2087 2022 1945
444 1865 1786 1713 1649 1596 1550 1512 1478 1448 1422 1401
:

The standard can be downloaded here: “JCAMP-DX: A Standard Form for the Exchange of Infrared Spectra in Computer Readable Form“, ROBERT S. McDONALD and PAUL A. WILKS, JR., Appl. Spectrosc. 42(1), pp151-162, 1988

What is NIR-Spectroscopy? (simple explanation, simply explained)

In the most cases a simple Halogen lamp emits light including the near infrared (NIR) spectrum (harmless radiation) to the sample/probe and the reflected light is measured. The light loses some energy on-and-in the sample depending on its physical and chemical (molecular) structure. The missing part of the light is treated as a fingerprint of the sample that is mathematically analyzed with prefabricated NIR calibration models (built with chemometric methods), based on trained known samples. That makes it possible to simultaneous analyze multiple physical- and chemical-properties (constituent, ingredient, analyte) within a few seconds and is non-destructive to samples.

The Ghost Calibrator

To explain our service in an other way, I use an analogy between a book and a calibration. Building good calibrations is like writing a good book (a bestseller). You can write in a foreign language (chemometrics) with a high sophisticated word-processor (the chemometric software) that has a grammar checker (an outlier detection). Due to the complexity of the language (chemometrics) and the difficulty of the chosen book topic (the data) and the incomplete automatic grammar checker, you can never be sure if the grammar is correct and may not lead to misunderstanding (bad prediction performance). So the best way is to let a native language speaker check and correct the text. In that way (the analogy), you can see us even as a ghostwriter (a ghost calibration developer, a ghostcalibrator) that helps you, writing the book (with long year experience, consolidated knowledge, time saving, a lot of benefit). The analogy fits very well, because you can define the topic of the book (with your data). Finally you own the calibration and you have the full insight in how it is done. You have it under full control.