Exploratory Analysis of South American Wines Using Artificial Intelligence

Area of Science:

• Analytical Chemistry
• Spectroscopy
• Machine Learning

Background:

• Geographical origin is a key factor in wine quality and authenticity.
• Accurate classification of wine by origin requires reliable analytical data and robust statistical methods.

Purpose of the Study:

• To apply microwave-induced plasma optical emission spectrometry (MIP OES) for multielement determination in South American wines.
• To evaluate the efficacy of machine learning algorithms (logistic regression, support vector machine, decision tree) in differentiating wine samples based on their geographical origin.

Main Methods:

• Acid digestion of 47 Brazilian and Argentinian red wine samples.
• Multielement analysis using microwave-induced plasma optical emission spectrometry (MIP OES).
• Exploratory data analysis and classification using logistic regression, support vector machine, and decision tree algorithms.

Main Results:

• Quantification limits (mg L⁻¹) were established for P, B, K, Mn, Cr, and Al.
• Accurate classification of all wine samples according to their geographical origin was achieved.
• Elemental concentration ranges for Al, Cr, Mn, P, B, Pb, Na, and K were determined.

Conclusions:

• MIP OES is a suitable technique for multielement analysis in wine.
• Machine learning algorithms can effectively differentiate South American red wines by region of origin using elemental composition data.