The Design And Implementation Of News And Information Processing System Based On The Semantics

Current methods to process news information mainly include automatic gatheringand manual annotation. In this procedure, the title and contents of the news are firstobtained, and then key words and topic labels are manually extracted. This method iseasy to practice; meanwhile it provides the ability to correlate news information by theirkey words. Most existing news information process systems adopt such traditionalmeans in their daily work. In the process, both the automatic gathering and manualanalysis of news information are required. The automatically gathered data are input tothe system, which need to be processed and selected by information experts. Theycarefully choose key words to best describe the given news from a certain topic wordset. At last the preprocessed data are collected by information summarization experts tofinish the whole procedure. This complete process can effectively categorize differenttypes of news information data, and can achieve high accuracy and commonunderstanding dependent on the solid support by its perfect topic word library.In this thesis, semantic technology was utilized to realize automatic data extractionand knowledge acquisition of news information. The reasonable semantic process wasas follows: representing knowledge as ontology, storing extracted knowledge asontology model, and using Allegro Graph to perform knowledge storage and knowledgequery. To implement the hot-spot discovery function in the system, a way of incrementalclustering combined with semantic knowledge base for realization was introduced in thethesis. In order to provide an efficient news information gathering pattern to replacecurrent information organization and visualization, we performed research of intelligentprocessing, organizing, querying, pushing, and complex analysis and data mining suchas hot-spot digging, trend-forecast, information-correlation to swiftly, accurately locateuseful information out of myriad pieces of unrelated text data. And only in this way canwe earnestly help news information processing or decision-making workers to gainfeasible strategy support, meanwhile promote the timeliness of information processing,and finally to convert information advantage into commercial advantage or decisionadvantage. The semantic-based intelligent information processing technology in our studyrepresents key objective that needs to be further investigated in scientific fields ofinformation fusion and myriad information data analysis. Ideas proposed in the thesisprovide key technology support for developing corresponding information processingsoftware system with equivalent intelligent processing ability.