Primary Research On Text Knowledge Atuo-Extraction Technique In The Aviation Domain

Now CBT development level is higher every day, correlative technology is more mature. In the development process, CBT need knowledge database. So the text language knowledge automatism extraction becomes one of the key technique in the CBT technology development.In order to automatic extract text knowledge, The paper research and carry out automatic extraction from aviation domain using information extraction technology and natural language processing. The paper limits the range of the text is aviation domain, knowledge form uses text language. The paper concludes the typical knowledge and mode through analysis and summarize by handwork. And the paper obtains pattern rule and forms rule database through regular expression.The paper use word membership and sentence membership of knowledge to filter the sentences from wide matching. Accounting Word membership is based on aviation material and Peoples'daily material. First the paper get word segmenting to the two material, and count word membership. Through word membership sentence membership is counted. If the knowledge appears two or more, it is conflicted when to save to knowledge database. The solution to the problem is by using VSM counting similitude, which select the max from.Experiments indicate that, the knowledge automatic extraction system is successful. The precision is 76.51 percent, and the recall is 83.78 percent. Finally, the conclusion and expectation are is put forward.