Skip to main content
. 2020 Aug 12;22(8):e17478. doi: 10.2196/17478

Table 5.

Description of deep learning classifiers, target, and parameter settings used in the experiments.

Deep learning classifiers Targets Parameter values
Vaping-related word vectors

CNNa Relevance max_features: 166,395, embed_size: 300, max_len: 75, optimizer: rmsprop, filters: 100, kernel_size: 1, epochs: 5, batch_size: 16

LSTMb Relevance max_features: 166,395, embed_size: 300, max_len: 75, optimizer: adam, epochs: 10, batch_size: 16

LSTM-CNN Relevance max_features: 166,395, embed_size: 300, max_len: 75, optimizer: adam, filters: 50, kernel_size: 2, epochs: 10, batch_size: 16

BiLSTMc Relevance max_features: 166,395, embed_size: 300, max_len: 75, optimizer: adam, epochs: 10, batch_size: 16

CNN Commercial max_features: 166,395, embed_size: 300, max_len: 75, optimizer: adam, filters: 100, kernel_size: 2, epochs: 10, batch_size: 16

LSTM Commercial max_features: 166,395, embed_size: 300, max_len: 75, optimizer: rmsprop, epochs: 5, batch_size: 32

LSTM-CNN Commercial max_features: 166,395, embed_size: 300, max_len: 75, optimizer: rmsprop, filters: 75, kernel_size: 2, epochs: 5, batch_size: 16

BiLSTM Commercial max_features: 166,395, embed_size: 300, max_len: 75, optimizer: adam, epochs: 5, batch_size: 64

CNN Sentiment max_features: 166,395, embed_size: 300, max_len: 75, optimizer: rmsprop, filters: 100, kernel_size: 2, epochs: 10, batch_size: 32

LSTM Sentiment max_features: 166,395, embed_size: 300, max_len: 75, optimizer: adam, epochs: 5, batch_size: 64

LSTM-CNN Sentiment max_features: 166,395, embed_size: 300, max_len: 75, optimizer: adam, filters: 75, kernel_size: 3, epochs: 5, batch_size: 64

BiLSTM Sentiment max_features: 166,395, embed_size: 300, max_len: 75, optimizer: rmsprop, epochs: 5, batch_size: 32
Global Vectors for Word Representation word vectors

CNN Relevance max_features: 15,890, embed_size: 200, max_len: 75, optimizer: adam, filters: 100, kernel_size: 2, epochs: 10, batch_size: 16

LSTM Relevance max_features: 15,890, embed_size: 200, max_len: 75, optimizer: adam, epochs: 5, batch_size: 32

LSTM-CNN Relevance max_features: 15,890, embed_size: 200, max_len: 75, optimizer: adam, filters: 50, kernel_size: 2, epochs: 10, batch_size: 16

BiLSTM Relevance max_features: 15,890, embed_size: 200, max_len: 75, optimizer: adam, epochs: 5, batch_size: 64

CNN Commercial max_features: 10,842, embed_size: 200, max_len: 75, optimizer: rmsprop, filters: 50, kernel_size: 2, epochs: 5, batch_size: 16

LSTM Commercial max_features: 10,842, embed_size: 200, max_len: 75, optimizer: adam, epochs: 5, batch_size: 16

LSTM-CNN Commercial max_features: 10,842, embed_size: 200, max_len: 75, optimizer: adam, filters: 75, kernel_size: 2, epochs: 5, batch_size: 32

BiLSTM Commercial max_features: 10,842, embed_size: 200, max_len: 75, optimizer: adam, epochs: 5, batch_size: 64

CNN Sentiment max_features: 7979, embed_size: 200, max_len: 75, optimizer: rmsprop, filters: 100, kernel_size: 3, epochs: 5, batch_size: 64

LSTM Sentiment max_features: 7979, embed_size: 200, max_len: 75, optimizer: adam, epochs: 5, batch_size: 32

LSTM-CNN Sentiment max_features: 7979, embed_size: 200, max_len: 75, optimizer: rmsprop, filters: 75, kernel_size: 1, epochs: 10, batch_size: 64

BiLSTM Sentiment max_features: 7979, embed_size: 200, max_len: 75, optimizer: adam, epochs: 5, batch_size: 32

aCNN: convolutional neural network.

bLSTM: long short-term memory.

cBiLSTM: bidirectional long short-term memory.