Skip to main content

Table 1 Sample distribution for the binary and multiclass versions of the collected dataset

From: On the utilization of deep and ensemble learning to detect milk adulteration

Multiclass # Samples Binary # Samples
Raw 2470 Raw 2470
Sucrose 486 Adulterated 2376
Formaldehyde 485   
Starch 480   
Peroxide 465   
Bicarbonate 460   
  1. The full dataset has 4846 samples. The raw class is ≈50% of the dataset. Each adulterant class is ≈10% on multiclass, while the binary considers these samples as one class (≈50%)