Construction of early-warning model for secondary and derivative events of public health emergencies based on event evolutionary graph
To predict the evolution situation of public health emergencies,an early-warning model for the secondary and de-rivative events of public health emergencies based on the event evolutionary graph was proposed.The two-layer ontology was utilized to achieve the structured scenario representation,the pre-training model and deep neural network were comprehensive-ly applied to realize the event extraction,then the pattern matching approach was employed to extract the logical relationship of events,and the word embedding models and clustering algorithms were integrated to achieve the event generalization,there-by realizing the construction of event evolutionary graph,the similarity calculation and logical prediction were performed to a-chieve the early-warning for the secondary and derivative events.Moreover,the empirical research was conducted by emplo-ying"Influenza A virus"and"M.Pneumonia"events as examples.The results show that the"Influenza A virus"event evo-lutionary graph can clearly display the logical correlation between this event and the related events,and based on this graph,the secondary and derivative event early-warning for the"M.Pneumonia"event can be achieved.The research results can provide reference for predicting the secondary and derivative events of public health emergencies.
public health emergenciesevent evolutionary graphsecondary and derivative eventssimilarity calculationinterpretable early-warning
NETL
NSTL
万方数据
