Skip to main content

Table 1 Machine learning model comparison (values in the boxes are prediction accuracies on test data; higher values are better) (* Demonstrated deep learning model was a feed forward artificial neural network with three hidden layers)

From: Machine learning analysis of microbial flow cytometry data from nanoparticles, antibiotics and carbon sources perturbed anaerobic microbiomes

Putative Groups

Gradient Boosting

Naïve Bayes

Distributed Random Forests

Deep Learning*

Acetogens

41.87%

63.87%

18.00%

52.67%

Acidogens

91.20%

97.07%

53.07%

99.73%

Hydrolyzers

65.60%

67.20%

10.67%

57.07%

Methanogens

85.17%

44.75%

89.33%

76.83%

Overall

71.26%

60.44%

53.55%

70.55%