DC Field | Value | Language |
---|---|---|
dc.contributor.advisor | Segev, Aviv | - |
dc.contributor.advisor | 세게브, 아빕 | - |
dc.contributor.author | Lee, Jae-Hwa | - |
dc.contributor.author | 이재화 | - |
dc.date.accessioned | 2011-12-14T04:27:01Z | - |
dc.date.available | 2011-12-14T04:27:01Z | - |
dc.date.issued | 2011 | - |
dc.identifier.uri | http://library.kaist.ac.kr/search/detail/view.do?bibCtrlNo=467973&flag=dissertation | - |
dc.identifier.uri | http://hdl.handle.net/10203/41749 | - |
dc.description | 학위논문(석사) - 한국과학기술원 : 지식서비스공학과, 2011.2, [ vi, 41 p. ] | - |
dc.description.abstract | Maps such as concept maps and knowledge maps are often used as learning materials. These node-link tools (or maps) have nodes and links, nodes as key concepts, links as relationship between key concepts. From a map, user can recognize what are important concepts and what relationships exist between them. Node-link tools have some advantages over texts. First, minimum texts are used in maps, so that users can easily search for words, phrases, and ideas.[34] Another advantage is that learners learning from maps can have several learning paths, unlike learning from texts which only can have one path of learning, either top-to-bottom or left-to-right.[22] However, there are also concerns of using maps. One is if a map is big, users of the map can be overwhelmed by the size of it, map shock.[8] Another is that maps can have poor visual configuration when they fail to use node proximity to signal semantic similarity.[42] In order to build concept or knowledge maps, domain experts are needed who understand the topic well. It costs very much, and it is hardly possible for people to have one when they need. In this study, a trial of automatically building a domain knowledge map was made with text mining techniques. From a set of documents which are about a specific topic, keywords were extracted using tf/idf algorithm.[38] For every pair of keywords, the number of appearance in a sentence and the number of words in a sentence were considered to calculate relation rankings. Based on the two information, a domain knowledge map (K-map) was built. K-map is a domain knowledge map without labels on links. If it had labels for links, it would look much like a concept map. However, extracting words for those labels are not covered in this study, and left as future work. Instead of putting labels for relations, K-map shows all the sentences that contain the two keywords which are placed at both ends of the relation chosen. Two experiments were conducted with K-map. One is ext... | eng |
dc.language | eng | - |
dc.publisher | 한국과학기술원 | - |
dc.subject | text mining | - |
dc.subject | learning | - |
dc.subject | domain knowldege map | - |
dc.subject | K-map | - |
dc.subject | K-map Tools | - |
dc.subject | K-map Tools | - |
dc.subject | 텍스트 마이닝 | - |
dc.subject | 학습 | - |
dc.subject | 도메인 지식 맵 | - |
dc.subject | K-map | - |
dc.title | Creating knowledge maps for learning using text mining | - |
dc.title.alternative | 텍스트 마이닝 기법을 이용한 학습 지식맵의 생성 | - |
dc.type | Thesis(Master) | - |
dc.identifier.CNRN | 467973/325007 | - |
dc.description.department | 한국과학기술원 : 지식서비스공학과, | - |
dc.identifier.uid | 020093902 | - |
dc.contributor.localauthor | Segev, Aviv | - |
dc.contributor.localauthor | 세게브, 아빕 | - |
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.