Fig. 2From: On the utilization of deep and ensemble learning to detect milk adulterationBoxplot of the component features from the dataset, analyzed by ensemble learners. The plots show scales and variation of each feature. Cells, QValue and MUN have significantly different scales and were plotted separately from the other variables. The Cells plot show that most samples has somatic cells counting from ≈150 to ≈450. The QValue plot shows that most samples on the dataset were considered of high quality (≈87 to ≈93). The FrzPoint plot shows that all samples has a freezing point at just above zero. All the other plots consider specific component concentrations and they show low variationBack to article page