Table 3 The optimized parameter settings of the LSTM network

Word_dim

200

Token embedding dimension

Word_LSTM_dim

100

Token size in LSTM hidden layer

Word_bidirectional

TRUE

Using Bi-LSTM

Word Embedding

TRUE

Using word embedding

CRF

TRUE

Using CRF

Dropout

1

Input droupout

Learning method

SGD

SGD Adadelta Adam

Abbreviation

TRUE

Using Ab3P