The Deployment and Analysis of Textual Features on Chinese Teenagers Semantic Space

碩士 === 國立臺中教育大學 === 特殊教育學系碩士班 === 103 === The purposes of this study were to build Chinese teenagers semantic space, verified by semantic correlations between words and words, sentences and sentences, correlating in comparing with human scoring. And the use of Chinese Coh-Metrix analysis of develop...

Full description

Bibliographic Details
Main Authors: BAI,JIA-YU, 白嘉郁
Other Authors: LIAO,CHEN-HUEI
Format: Others
Language:zh-TW
Published: 2015
Online Access:http://ndltd.ncl.edu.tw/handle/5s548k
Description
Summary:碩士 === 國立臺中教育大學 === 特殊教育學系碩士班 === 103 === The purposes of this study were to build Chinese teenagers semantic space, verified by semantic correlations between words and words, sentences and sentences, correlating in comparing with human scoring. And the use of Chinese Coh-Metrix analysis of development indicators to extract the main characteristics,analysis the text features among various subjects in junior high school textbook.Finally, the text features predictive text suitable grade. The results are as follows: 1. The Chinese teenagers semantic space contains 706 article and a total of 285820 number of words. Category contains different areas of Chinese, social andscience. Its characteristic is the maximum number of words on the second grade, the maximum number of words in science areas. 2. In the aspect of correlations between words and words, sentences and sentences, Chinese teenagers semantic space was middle correlated in human scoring. The situation in junior high school to use, the effect is better than 1--9 grades semantic space, children Chinese semantic space and adults semantic space. 3. The analysis of each subject text feature section, the results show that high Narrativity and low Referential Cohesion in Chinese; high Vocabulary Complexity and low Narrativity in science; high Vocabulary Complexity and low Syntactic Coherence in social. 4. The variance explained by the 6 text characteristics was 62.3%. The best 3 predictors were Vocabulary Complexity, Referential Cohesion, and Narrativity.