Incremental model learning for spectroscopy-based food analysis

Abstract

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.

Challenges in the Use of AI-Driven Non-Destructive Spectroscopic Tools for Rapid Food Analysis

“…Inherent to this issue, the use of batch learning methods requiring full recalibration after the addition of each new sample is inappropriate for the building of these universal chemometric models. Hence, there is a growing demand for accurate open-source chemometric tools to allow for the evolution or incremental learning of existing models and, at the same time, minimise computational and spatial costs [ 85 ]. In parallel, privacy and ethical concerns related to data sharing and user rights may hinder the development of large-scale general models when different proprietary databases are distributed across different institutions, private organisations, or countries.…”

Automatic food and beverage authentication and adulteration detection by classification hybrid fusion

“…In particular, adapting existing one-class food classifiers to new food samples that may originate from new harvest years or test samples measured on a different or repaired instrument or under new measurement conditions. A few studies have been performed for classification of food products, [33][34][35][36][37][38] but additional work is needed.…”

Continuous statistical modelling in characterisation of complex hydrocolloid mixtures using near infrared spectroscopy

“…Moreover, a differentiation of the carrageenan+gum+excipient samples between carrageenan+gum+KCl and carrageenan+gum+MD samples is also noticeable. Regarding the testing phase, before the same data pre-treatment, linear interpolation was applied to the test spectra in order to get the desirable number of variables (n=1050) since they were acquired with a spectrometer with different spectral resolution [21,22] to reduce overfitting to the NIR calibration set.…”