Key-Concept Extraction for Concept Map Construction
| Autor: | Jian-Ren Chen, 陳建任 |
|---|---|
| Rok vydání: | 2014 |
| Druh dokumentu: | 學位論文 ; thesis |
| Popis: | 102 The goal of automatic construction of concept maps is to solve “English reading” problems by extracting important concepts of an article. In order to solve the problems in reading comprehension, keywords and keyphrases can help learners understand main ideas of the article. The method in this study essentially searches for the key concepts of an article by using statistical data, semantics information, syntax structure and constructing concept map to assist learners for improving reading in English. The concept map is divided into three layers. The first layer presents the central idea of the article. The second layer shows the main idea of each paragraph. The third layer consists of the supporting ideas of each paragraph. The statistical data are used such as term frequency, article title, semantic role labeling, lexical chains and hybrid methods to extract keywords. Then the keywords are further used to extract the keyphrases by regular expression and named entity recognition. In the experiments, we compare extracted keyphrases from benchmark datasets with the keyphrases assigned by human. Keywords：Concept map, Keyphrase extraction, Named entity recognition, Automatic semantic labeling, Lexical chain |
| Databáze: | Networked Digital Library of Theses & Dissertations |
| Externí odkaz: |
|