An Artificial Intelligence Approach for Italian EVOO Origin Traceability through an Open Source IoT Spectrometer

Area of Science:

• Food Science
• Analytical Chemistry
• Spectroscopy

Background:

• Extra virgin olive oil (EVOO) is a key component of the Mediterranean diet, with its quality and origin impacting consumer trust and value.
• Ensuring the authenticity of EVOO's geographical origin is critical due to its high market value and susceptibility to fraud.
• Reliable analytical tools are needed to authenticate EVOO geographical provenance.

Purpose of the Study:

• To evaluate the efficacy of an open-source visible and near-infrared (VIS-NIR) spectrophotometer, coupled with a dedicated app, for determining the geographical origin of EVOO.
• To demonstrate a cost-effective and accessible method for EVOO authentication.

Main Methods:

• Analysis of 67 Italian and 25 foreign EVOO samples using a VIS-NIR spectrophotometer.
• Processing spectral data with artificial intelligence algorithms, including multivariate analysis of variance (MANOVA) and artificial neural network (ANN) modeling.
• Identification of key spectral wavelength ranges associated with origin determination.

Main Results:

• MANOVA revealed significant differences (p < 0.001) in VIS-NIR spectral data between Italian and foreign EVOO samples.
• The ANN model achieved a 94.6% correct classification rate on an external test set.
• Significant spectral ranges for origin determination were identified between 308-373 nm and 594-605 nm, linked to phenolic compounds, carotenoids, and chlorophylls.

Conclusions:

• The VIS-NIR spectrophotometer and AI approach provide an efficient and reliable method for assessing EVOO geographical origin.
• This technology can help combat fraud and adulteration, safeguarding the integrity of "Made in Italy" EVOO.
• The findings support the use of accessible spectroscopic tools for food authentication and quality control.