Journal of Healthcare Engineering / 2017 / Article / Tab 1

Research Article

Combination of Deep Recurrent Neural Networks and Conditional Random Fields for Extracting Adverse Drug Reactions from User Reviews

Table 1

Results of the proposed models and baseline methods.


Baseline CRF0.62540.59720.61100.81450.75390.7521
Feature-rich CRF0.67260.65320.66280.83030.76460.7622
1-layer LSTM0.57980.65870.61670.81210.80650.7809
2-layer LSTM0.63620.70440.66860.80900.84950.8005
3-layer LSTM0.65880.70220.67980.82470.83230.7997
4-layer LSTM0.66890.70930.68850.82550.82800.8000
1-layer GRU0.58620.67720.62840.79950.83680.7900
2-layer GRU0.63840.70930.67200.81650.83380.8002
3-layer GRU0.66750.71910.69230.81510.83730.8009
4-layer GRU0.65650.72620.68960.80060.86650.8033
2-layer LSTM + CRF0.69470.69730.69600.81910.81610.7872
2-layer LSTM + CNN + CRF0.68090.70390.69220.80830.84880.7978
3-layer LSTM + CNN + CRF0.68680.70660.69650.82700.84880.8115
3-layer GRU + CNN + CRF0.70480.70820.70650.82190.83110.7978

We are committed to sharing findings related to COVID-19 as quickly and safely as possible. Any author submitting a COVID-19 paper should notify us at to ensure their research is fast-tracked and made available on a preprint server as soon as possible. We will be providing unlimited waivers of publication charges for accepted articles related to COVID-19. Sign up here as a reviewer to help fast-track new submissions.