Applying Ontology-based System to Construct the Autonomous Categorization for Questions

• Main Authors: Chang, Shiang Yu, 張翔宇
• Other Authors: DAY, Jiann-Yun
• Format: Others
• Language: zh-TW
• Published: 2003
• Online Access: http://ndltd.ncl.edu.tw/handle/66242692501212747803

碩士 === 國立臺灣師範大學 === 工業教育研究所 === 92 === In order to manage and use knowledge efficiently, the technologies of ontology are widely applied to various kinds of domain knowledge. This thesis proposes an automatic ontology construction approach that can help knowledge manager efficiently construct a domain knowledge. we propose a information classification model based on the constructed ontology. And the system can simplify tasks about the categorization of questions from any source; this is a marvelous tool for educators or scholars. As for learners, it assists them to comprehend each question’s ability and enhances learning efficiency. For question developers, It is used to analyze the quality of questions. The first feature of this thesis is to utilize an object-oriented approach to represent the structure of ontology, called object-oriented ontology. Second, the automatic ontology construction approach is presented. We embed CKIP system to carry out the Chinese natural language processing including part-of-speech tagging, Chinese-Term analysis and Chinese-Term feature selection by calculate the term weighting from operational item bank of PreMOUS Word certificate. Third, the concept clustering mechanisms for domain ontology automatic construction based on Neuron technology and data mining technology. This study consists PreMOUS Word (MOCC Word Core) certificate model, developed by “The Chinese Computer Education Association” (CCEA). The categorizing is applied for questions in PreMOUS Word, which creates a possibility of 90% accuracy rate. Applying this system raises the speed effectively of categorizing, clarify the architectonic and association between the concepts in the test bank; therefore, it is the target study for item bank’s quality control.