Skip to main content

View full-text article in PMC

. 2020 Aug 12;22(8):e17478. doi: 10.2196/17478

Table 7.

Performance of commercial classifiers.

Classifiers	Area under the receiver operating characteristic curve (95% CI)	Precision	Recall	F1
Logistic regression	0.98 (0.95-0.99)	0.93	0.83	0.96
Random forest	0.97 (0.96-0.99)	0.95	0.82	0.97
Support vector machine	0.98 (0.91-0.99)	0.92	0.86	0.92
Naive Bayes	0.96 (0.94-0.99)	0.83	0.89	0.92
CNN^a (vaping-related word vectors)	0.98 (0.96-0.99)	0.93	0.75	0.94
LSTM^b (vaping-related word vectors)	0.97 (0.95-0.99)	0.88	0.81	0.94
LSTM-CNN (vaping-related word vectors)	0.97 (0.94-0.99)	0.92	0.85	0.94
BiLSTM^c (vaping-related word vectors)	0.98 (0.96-0.99)	0.84	0.87	0.95
CNN (GloVe^d word vectors)	0.99 (0.98-0.99)	0.93	0.89	0.98
LSTM (GloVe word vectors)	0.99 (0.98-0.99)	0.89	0.94	0.98
LSTM-CNN (GloVe word vectors)	0.99 (0.98-0.99)	0.86	0.96	0.99
BiLSTM (GloVe word vectors)	0.99 (0.98-0.99)	0.97	0.88	0.98

^aCNN: convolutional neural network.

^bLSTM: long short-term memory.

^cBiLSTM: bidirectional long short-term memory.

^dGloVe: Global Vectors for Word Representation.