Skip to main content

Table 3 Summary of arthropod machine learning model performance

From: Detecting false positive sequence homology: a machine learning approach

 

OrthoDB Arthropod EQUAL

OrthoDB Arthropod PROP

Algorithm

Validation

Testing

Validation

Testing

Neural Network

97.1815 %

96.8153 %

97.5452 %

96.5423 %

Suppor Vector Machine (SVM)

89.1351 %

88.0801 %

88.0668 %

88.2621 %

Random Forest

98.1362 %

95.9054 %

97.8748 %

95.5414 %

Naive Bayes

53.0628 %

52.5023 %

61.2229 %

60.3276 %

Logistic Regression

96.5905 %

97.2702 %

96.3064 %

96.3603 %

Meta-Classifier w/o Logistic Regression

98.5112 %

98.3621 %

98.5907 %

96.8153 %

Meta-Classifier w/ Logistic Regression

98.6362 %

97.7252 %

98.5680 %

97.5432 %

  1. This table shows the performance of each of the different learning algorithms that were trained, validated, and tested with the OrthoDB arthropod gene clusters