Study On Tourism Emergency Information Integration System Based On Agent

In order to change the current security situation of tourism, it is urgent to improve greatly the prediction capability with tourism emergencies to adapt to the situation of national tourism's economic development and satisfy domestic and foreign tourists with the information requirement to tourism emergencies. This thesis makes research on agent based information integration system of tourism enmergency. Main works are as follows:(1)Select data source and set data format: provide the method that we can express both structured and no-structured tourism enmergency data by XML file, and introduce the DOM technology which is used to combine all data from different information source.(2)Put forward a information extraction method based on event framework which can extract time, space and harm information from events. We define four kind of frameworks for disaster event, constituted extraction rules based on corpus, completed extraction of time, space and harm information in disaster warning report.(3)Bring forward a text segmentation algorithm based on sentence clustering technology. We decide the features which are used in the sentence clustering process, compare the difference of the sentence similarity calculation method between the VSM based technology and the LSI based technology, present a new sentence clustering algorithm that combines the K-centers clustering and hiberarchy clustering algorithm and compare with the conventional hiberarchy clustering algorithm. The results show that the improved sentence clustering algorithm has a higher accuracy of text segmentation.(4)Construct an agent based information integration platform: we present the architecture of agent based information integration system, develope the prototype system of tourism emergency information integration system based on Jadex platform. Experiment indicates that the prototype system can integrate successfully information from different sources.According to the characteristic of tourism related information distributing in heterogeneous data source, this thesis integrates information of various information sources by using information integration technology, and lays the foundation for the next data mining and predictive analysis.