| In CBT courseware development platform, many function and application are supported by multimedia knowledge database. The text language knowledge is the most basic knowledge form, so text knowledge automatism extraction becomes the key technique in the CBT courseware development.The paper mostly introduces how the knowledge extraction system is designed and realize in the CBT courseware development. The system is input with natural language which is described in the electronic text form. By pattern matching, it takes out the knowledge which the user is interested with and is regulated in advance. Then the knowledge is organized and structured. In the end, it is storged in the database in a normative form. The form is used easily to consequence. By this way, the system realizes the text knowledge's auto-search and auto-storage in the structure of the multimedia knowledge database in the CBT courseware development platform.The paper limits the range of the text which is wanted to be extracted in the teaching material of the civil aviation. By gathering many sentences which is used to describe the typical knowledge, the exploder team constructs the training set. Through analysis and summarize, all kinds of the language characters is gained. Then the characters and the conclusion are described in the Production rule form. In other words, the rule set is constructed by handwork. In these bases, a mutual knowledge extraction system is programmed. The original knowledge which is described in natural language in electronic text is disposed by filtrating, semantic block segmenting, word segmenting, syntax and semantic analyzing, pattern matching. In the end it is expressed in the semantic network and storaged in the multimedia knowledge database. |
PDF Full Text Request