Ci spiace, ma questo articolo è disponibile soltanto in English.
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.
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).
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.
We help you to find the optimal settings for higher NIR accuracy and reliability. You can build your own custom NIR calibration model with this valuable settings. We offer a quantitative NIR Calibration development and optimization service. New: White Paper about the details, what’s behind.
Increase Your Profit with optimized NIR Accuracy
- going closer to your product specification limits and maximize profitability
- optimizing your models yield to process optimization and optimizing productivity
- compete against other NIR vendors in a feasibility study (NIR salesman)
- compatible with any NIR vendor
- no installation, no learning
- quantitative NIR Calibration Development as a Service
- help users avoid common pitfalls of method development
- before you validate and approve your solution for use in production process:
- check if a better calibration can be found,
- compare your calibration with other experts solutions.
- no cumbersome trial-and-error modeling steps
- calculation time is spent on our high performance infrastructure
- fast results, developed calibrations within days
- fix costs, depends only on data size (not hourly rate for service)
- huge saving in method development costs
- easy to plan
There are many published standards and guidelines (protocols) available for developing NIR calibrations from Standards Consortium such as ASTM, EMEA, ICH, IUPAC, ISO, USP, PASG etc. as well as many good recommendations and guidelines found in various textbooks and papers.
The difficulty with so many ‘Protocols’ for the NIR user is to have them all available and in their thought processes during calibration work and in addition to execute, check and challenge all calibrations generated manually. This is time consuming and sometimes boring repetitive work.
To simplify this for the person generating the NIR Calibrations, we have collected the good practices protocols and integrated them into our service that automates the calibration building and evaluation procedures.
to part 2
5. Calibration Precision
What do you believe, can NIR calibration models be more precise than reference values?
6. Calibration Maintenance
How often do you update your quantitative calibrations per year?
7. Quantitative Calibrations
How many quantitative (%) calibrations do you have in use?
8. Quantitative Parameters
In all your quantitative calibrations, how many parameters (properties) you have in total?
9. Qualitative Calibrations
How many qualitative (identification) calibrations do you have in use?
Please vote and see the assessments below.