Classification and evaluation of cow’s milk according to the place of origin based on compositional parameters utilizing Machine Learning Techniques
Theodoros Markopoulos , Sotirios Papadopoulos , Stavros Kontakos , Alexandros Tsoupras
AIMS Agriculture and Food ›› 2025, Vol. 10 ›› Issue (4) : 989 -1003.
The increasing expectations regarding food quality in recent years have prompted a detailed investigation into the factors that define this concept. Several of these factors are strongly shaped by the production method, while in other cases, compositional characteristics are influenced by the production region itself. This creates a distinctive relationship between the product’s basic characteristics and its place of origin, which can be leveraged to ensure consistently high discrimination with minimal variation. In this study, we examined the compositional parameters of 84,425 cow’s milk samples collected from individual farms of neighboring regions in northern Greece, aiming to assess their relationship with the geographical area of production. Four supervised Machine Learning classification methods-k-Nearest Neighbors, Decision Tree, Random Forests, and Support Vector Machines-were employed, all suitable for Big Data analysis. The findings indicate that all four methods consistently classify the milk samples meaningfully only for two of the four regions, specifically the prefectures of Serres and Xanthi. As anticipated, the Random Forest algorithm achieved the strongest classification performance among the tested techniques.
dairy / implementation / machine learning / classification / certification / production
| [1] |
Steenkamp JBEM (1990) Conceptual model of the quality perception process. J Bus Res 21: 309-333. https://doi.org/10.1016/0148-2963(90)90019-A |
| [2] |
|
| [3] |
|
| [4] |
Markopoulos T, Stougiannidou D, Kontakos S, |
| [5] |
|
| [6] |
|
| [7] |
|
| [8] |
|
| [9] |
|
| [10] |
|
| [11] |
|
| [12] |
|
| [13] |
|
| [14] |
|
| [15] |
|
| [16] |
|
| [17] |
|
| [18] |
R Development Team (2008) R Language definition. Comprehensive R Archive Network. Available from: http://cran.rproject.org. |
| [19] |
Breiman L (2001) Random Forests. Mach Learn 45: 5-32. https://doi.org/10.1023/A:1010933404324 |
| [20] |
|
| [21] |
|
| [22] |
|
| [23] |
|
| [24] |
Quinlan RJ (1986) Induction of decision trees. Mach Learn 1: 81-106. https://doi.org/10.1023/A:1022643204877 |
| [25] |
|
| [26] |
|
| [27] |
|
| [28] |
|
| [29] |
FAO (2021) The nutrition and health potential of geographical indication foods. Rome. Available from: https://doi.org/10.4060/cb3913en. |
/
| 〈 |
|
〉 |