Research On Web Based Tourism Emergency Event Retrieval System

The research work in this dissertation is supported by the National Natural Science Foundation of China (60773112): the analysis of information pattern and study on prediction in tourism emergencies, and the Beijing Natural Science Foundation (4082021): the study of data mining and intelligent prediction in tourism emergencies, to help users and government officials retrieval travel emergencies in the web. This dissertation has done some research in web based tourism emergencies retrieval system. The main job and innovation are as follows:(1) Crawl and analyze the tourism emergency pages. Use spider technology to crawl tourism emergency pages from internet. Propose the block based algorithm and combine with web analytics technology to obtain key information such as the body text information to form local tourism emergencies collection.(2) Realize automated classification of tourism emergencies texts. Manually mark the crawled travel emergencies to form a training set. After the segmentation of training set, use the improved IG algorithm to accurately calculate the characteristics of every word, and apply sort algorithm to extract feature words set at last. Use IFIDF technology to build VSM vector space. Finally use SVM algorithm to build a classifier, and test the classifier performance. The classifier achieves good effect in our experiment.(3) Realize the extraction of key attributes of tourism emergencies. Use Chinese text information extraction technology based on HMM to learn the train set which is marked manually and finally build the extraction model.(4) Realize the development of web based tourism emergencies retrieval system prototype by using BOOL retrieval model to build the query interface.This dissertation combines Chinese text processing technology, Web retrieval technology and the application of tourism emergency retrieval to achieve a good travel emergency retrieval system with high rate of accuracy, speed and scalability. The system helps relevant staff grasp the information of tourism emergencies and provide strong technology support to the research afterwards.