This study examines the compositional differentiation of two Cucurbita species, C. maxima Duchesne and C. moschata Duchesne, to identify chemical markers relevant for their nutritional and functional potential. Multivariate statistical analysis, including principal component analysis (PCA), was applied to chromatographic, chemical, and antioxidant descriptors to visualize patterns of variability among samples. Classification artificial neural network (cANN) models were used to explore the potential of machine learning for sample differentiation, using integrated lipidomic, carotenoid, phenolic, and liquid chromatographic datasets, providing a multidimensional biochemical characterization of Cucurbita samples, achieving good classification within the analyzed dataset, reflecting the model’s capacity to describe the available data. The integration of chemometric and ANN approaches provides a framework for the compositional profiling and quality assessment of Cucurbita species, offering insights into their sustainable valorization as sources of bioactive compounds for food and nutraceutical applications while acknowledging the need for further validation on larger datasets.
Miljić et al. (Thu,) studied this question.