0.0

Yazarlar:
Ardiana Topi Dritan Topi
Yayın Yılı:
2025
Yayıncı:
Liberty Publications
Dil:
ISBN:
979-8-89695-034-9
Anahtar Kelime (AI):

6TH INTERNATIONAL ISTANBUL CURRENT SCIENTIFIC RESEARCH CONGRESS
APPLICATION OF MACHINE LEARNING FOR THE FOOD AUTHENTICATION, CASE OF ALBANIAN OLIVE OILS

Yazarlar:
Ardiana Topi Dritan Topi
Yayın Yılı:
2025
Yayıncı:
Liberty Publications
Dil:
ISBN:
979-8-89695-034-9
Özet:
:
Food authentication is a rapidly growing field of scientific research due to increasing public awareness of food quality and safety. It is the process that verifies that food complies with its labeled description. This may include the origin (species, geographical, or genetic), production method (conventional, organic, traditional procedures, free-range), or processing technologies (irradiation, freezing, microwave heating). This study focuses on validating olive oils from Albania by examining local and foreign cultivars along with their geographical origins. To improve the accuracy of origin predictions, essential data preprocessing steps, particularly normalization after isolating independent features from the target variable, are vital for distance-based algorithms like kNN. This approach enhances overall accuracy. Additionally, we assessed the performance metrics for various algorithms, including k-nearest Neighbors, Logistic Regression, and Support Vector Machines, alongside hyperparameter tuning techniques applied to the best-performing model. We can effectively trace their geographical and cultivar origins by utilizing supervised machine learning methods to classify Albanian Olive Oils (OO) based on their chemical composition. Our findings reveal an accuracy rate of 88.88%, limited by the present dataset; however, we plan to broaden the dataset in the future. The results predict the development of a comprehensive dataset for olive oils produced in the country. This will serve as a basis for future proposals on producing OO with Designated Origin (PDO) and Geographical Indication (PGI).