Incremental model learning for spectroscopy-based food analysis

•Subspace-based incremental learning for improving multivariate analysis models in the field of chemometrics of spectral data.•It allows sharing models between analytical laboratories without privacy issues.•It allows improving preexisting models with a small number of samples up to 20%.•It allows adapting a model to a new acquisition equipment with an improved in performance up to 60%.•It can be applied as a classification tool or as a data exploratory technique..

In this paper we propose the use of incremental learning for creating and improving multivariate analysis models in the field of chemometrics of spectral data. As main advantages, our proposed incremental subspace-based learning allows creating models faster, progressively improving previously created models and sharing them between laboratories and institutions without requiring transferring or disclosing individual spectra samples. In particular, our approach allows to improve the generalization and adaptability of previously generated models with a few new spectral samples to be applicable to real-world situations. The potential of our approach is demonstrated using vegetable oil type identification based on spectroscopic data as case study. Results show how incremental models maintain the accuracy of batch learning methodologies while reducing their computational cost and handicaps.