The Information Detection Method For Typhoon Event Process Based On Microblog Text

Typhoon will have a serious impact on natural ecology,social economy and even human sustainable development.Real-time detection of the evolution process of typhoon events has become an important basis and reference for disaster emergency response.At the age of big data,social media has become the main platform for people to create and share information,and is breaking down the time and space barriers of information dissemination.With its efficient update frequency,multi-source communication channels and extensive participation,social media has shown great potential in disaster management and has gradually developed into a new way to obtain information on typhoon events.However,due to the short-text nature of social media,it also has the characteristics of high information fragmentation,diverse forms of expression and multiple granularities of information.Distributed social media information is not only difficult to reflect the whole evolution of typhoon events,but also hinder users from detecting the process of typhoon events effectively.Drawing on the scientific theories and methods in the fields of disaster,geography and information science,this paper uses microblog texts as data sources,and sets the disaster mechanism of the interaction among pregnant environment,causing factor and sustaining body as the basis.Firstly,the information model and formal representation method for the typhoon event process are constructed.Then,the multi-type typhoon event information extraction method based on microblog text is proposed.Finally,the aggregation method considering multi-granularity and dynamic characteristics of typhoon event information is explored to realize the information detection of typhoon event process based on microblog text.The main contents and results of this paper include the following aspects:(1)Multi-level information modeling of typhoon eventsBased on the analysis of the components and dynamic characteristics of typhoon events,an information model of typhoon events(TEIM)is proposed from three levels:object,process and state.This model considers the multi-granularity of typhoon event information,and unifies the different feature information into each object.Besides,in order to highlight the evolutionary characteristics of typhoon events,the state is taken as the cross-section of the process in a specific spatio-temporal feature,and a dynamic description mechanism of “state-process” is proposed.This model not only emphasizes the features of different objects in the event,but also expresses the evolution process of typhoon events through the state sequences of multi-objects.(2)Formal representation of typhoon event information based on finite state machineAccording to the structure of the information model of typhoon events,the traditional finite state machine method is improved,and a finite state machine extension model(TEFSM)based on the information characteristics of typhoon events is proposed.In this model,the description ability of temporal and spatial features is added,and the representation method of hierarchical structure and process is extended.On this basis,the information on typhoon events is formally represented,and the standardized description of typhoon events,their components and the relations are realized.(3)Information extraction of typhoon events based on combined rule model and machine learningBecause of the different granularity and complex expression of typhoon event information in microblog text,and the lack of open tagged corpus on typhoon event,the supervised learning method is difficult to be transferred directly.By analyzing the descriptive characteristics of typhoon event information in microblog texts,this paper explores lexical rules of typhoon event information considering context,and constructs a combined rule model and machine learning method for typhoon event information recognition and classification.Through structured extraction of object names and feature information about typhoon events in microblog texts,it provides an information basis for aggregation of the typhoon event process.(4)Information aggregation for typhoon event process based on "objectstate"Following the semantic and logical relationship of the information model of typhoon events,the information aggregation mode for typhoon event process(TEPIA)is proposed.Through the hierarchical aggregation of “object-state-process”,this mode solves the problems of decentralization,spatial-temporal granularity diversification and disorder of typhoon information extracted from microblog texts.The aggregation results take into account the hierarchical structure and evolution characteristics of typhoon event information,reveal the multi-dimensional relationship among the information scattered in microblog texts,and provide an ordered information basis for exploring the development stage and situation of the typhoon event process.(5)Case studyThis paper uses Sina Weibo data to study the case of the typhoon Lekima event in2019.Through the evaluation and analysis of the case results,the feasibility of detecting the typhoon event process based on microblog text is verified.The study shows that compared to the fixed-interval information gathering and forecasting model in traditional meteorological monitoring,the temporal and spatial granularity of information in microblog texts is more changeable,which is good at reflecting the sudden changes of weather conditions at a small scale in time,it can also make up for the lack of meteorological monitoring for the perception of social-environmental changes.Moreover,the typhoon event information model proposed in this paper lays a theoretical foundation for the study,and the methods of information extraction,formalization and information aggregation can realize the deep mining and orderly integration of typhoon event information in the microblog texts,it provides the method support for the study of the information detection of the typhoon event process.