Predictions for COVID-19 Transmission Trend with Probabilistic FastText model

  • Siyue Song (Speaker)
  • Chen, T. (Contributor to Paper or Presentation)
  • Sean Knox (Speaker)
  • Grigoris Antoniou (Speaker)

Activity: Talk or presentation typesOral presentation


Currently, the COVID-19 is spreading into the whole world, there is a highly need to apply different methods to evaluate and predict the transmission trend of such disease. This work aimed to apply the Probabilistic FastText model to the COVID-19 open research dataset, with the comparison of other popular deep learning models, such as Text CNN, Bi-LSTM and transformer models. It is found the Probabilistic FastText has the outstanding performance (accuracy 89%). Finally, LDA, fuzzy C mean models can combine with PFT to generate attributes, which can be used to compose of fuzzy rules to provide interpretation of the results. The above results are provided for healthcare experts’ reference.
Period16 Jun 2022
Event titleHealTAC 2022: The 5th Healthcare Text Analytics Conference
Event typeConference
Conference number5
LocationVirtual, OnlineShow on map
Degree of RecognitionInternational