Skip to main content

Table 2 Results of Classification Algorithms

From: Automated deception detection of 911 call transcripts

Classification methods

Logistic model tree induction

Naïve Bayes

Random forest

Neural network

Discriminant analysis

 

Training

Cross-valid

Training

Cross-valid

Training

Cross-valid

Training

Cross-valid

Training

Cross-valid

Overall performance

98%

82%

90%

78%

100%

70%

82%

74%

96%

84%

Truth performance

100%

82%

88%

76%

100%

64%

84%

64%

96%

88%

Deception performance

96%

84%

92%

80%

100%

76%

80%

84%

96%

80%