Revolutionizing food integrity: Harnessing the power of ensemble learning algorithms

The term “ensemble learning” refers to a class of techniques in machine learning (ML) that blends manifold individual models, so-called base learners to improve the performance and enhance the robustness of the overall model. In the context of food integrity, where authenticity, safety and quality are crucial aspects, ensemble learning methods have become popular over the years. Due to the vast applicability of ensemble learning algorithms in food analysis, the objective of this review is to introduce the principles of ensemble learning to a broad audience in a simplified way without complex mathematical details and formulas. Moreover, typical applications of ensemble learning algorithms in various fields of food analyses are covered. As ensemble methods feature high predictive power, they are useful for building robust models in food integrity problems. Techniques based on ensemble learning have usually better prediction results in terms of accuracy, sensitivity, and specificity than classical ML methods. • Ensemble learning within the context of food integrity is discussed. • Parallel and sequential ensemble learning algorithms for food integrity are explained. • Bagging, and boosting strategies in ensemble learning are presented. • Various applications of ensemble learning in the field of food integrity are reported.