網路適性閱讀環境的建立與應用-子計畫V：潛在語意分析及概念構圖在文章摘要和理解評量的應用(III)
No Thumbnail Available
Date
2005-07-31
Authors
張國恩
宋曜廷
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
本計畫主要的目的為發展摘要寫作評
量系統。評量方式主要參考「潛在語
意分析」的技術，從教材及學生的摘
要作品中，找出關鍵詞做為評量的依
據。在應用潛在語意分析時我們採用
關鍵詞向量的比對，比對指標考慮到
向量間的COS、Entropy、向量長度等
參數來計算文件的相似程度。
The proposal aims to develop an evaluation system for summarization. LSA (Latent Semantic Analysis) approach is used in the evaluation. LSA first finds keywords from teaching materials and student’s summarization, and forms two keyword vectors. Based on the two keyword-vectors, LSA can calculate the similarity between them. The similarity index used in LSA may be COS, entropy, vector length and their combinations. In addition, the formula of summary efficiency is also considered in the evaluation.
The proposal aims to develop an evaluation system for summarization. LSA (Latent Semantic Analysis) approach is used in the evaluation. LSA first finds keywords from teaching materials and student’s summarization, and forms two keyword vectors. Based on the two keyword-vectors, LSA can calculate the similarity between them. The similarity index used in LSA may be COS, entropy, vector length and their combinations. In addition, the formula of summary efficiency is also considered in the evaluation.