Skip to main content

Table 3 Accuracy from simpler classifiers (Logistic Regression, Linear Regression, and Partial Least Squares) for binary and multiclass classifications that serve as baseline for our deep and ensemble learners

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

Dataset Classification Logistic Regression Linear Regression Partial Least Squares
90/10% Multiclass 0.5876 0.3155 0.3526
  Binary 0.7649 0.7959 0.7691
75/25% Multiclass 0.5693 0.3350 0.3267
  Binary 0.7583 0.7962 0.7739
50/50% Multiclass 0.5592 0.3281 0.3256
  Binary 0.7140 0.7920 0.7714
  1. All classifiers were evaluated with 3 pairs of training and test datasets randomly selected from our milk samples, identified by their proportion of training and test samples