Development of product identification models

To gain responsiveness and quality

Development of authentication and identification models

You are involved in the production or R&D industry and you are faced with the problems of identifying raw materials upon receipt or detecting non-conformities on finished products?

You want to

Develop

precise and robust models for authentication

Maintenance des calibrations spectroscopiques

Maintain

your existing identification models

Transfer

these models from one production line to another

All the measurements

All the applications

Data analysis methods for discrimination, classification and identification of products and fault detection are your solution. These “Pattern Recognition” models make the sorting of your raw materials more reliable upon receipt and qualify your finished products, to significantly reduce your production costs and batch release times.

Data used for the construction of product authentication models can have very diverse origins, such as optical spectral measurements (near infrared spectroscopy (SPIR-NIR), mid infrared (MIR – FTIR), Raman, UV, etc.) or still physico-chemical measurements from a wet chemistry laboratory, chromatographic or mass spectroscopy measurements (HPLC, GC, GC-MS, etc.)… All these parameters from different sources can be combined by multi-block methods in order to produce more precise identification models.

Thanks to our knowledge in Chemometrics and Machine Learning applied to data from spectroscopy, laboratory measurements or industrial processes, we assist you in the various stages of development, optimization and maintenance of identification models and authentication for precise and robust classifications.

Classification, discrimination, identification

Classification (PLS-DA, SIMCA, Support Vector Machines (SVM), Decision trees (CART, RF, XGBoost…)
Validation and model robustness
Transfer of calibrations (standardization (PDS, etc.), orthogonalization methods, etc.)
Model update
Implementation of models for routine use

Machine Learning methods

Software implementation

We develop your calibrations with all the Chemometrics software on the market (Unscrambler^®, SOLO^® or PLS_Toolbox^® in the MATLAB^® environment, SIMCA^®, etc.), as well as those provided by spectroscopy equipment manufacturers (OPUS, WinISI, NIRCal, TQ Analyst, etc…), but also free software (R, Python…).

Machine Learning software

Programming software for data analysis

Software from spectroscopic manufacturers

All the software

Our expertise for the analysis of your data

With more than 15 years of experience in data analysis (chemometrics and Machine Learning), in particular applied to measurements from spectrometers, laboratory data or processes, the experts of our teams support you at each stage of your projects.

Ondalys services