An Interactive Tool for English Reading Comprehension based on Partial Parsing and Construction Grammar

• Main Authors: Chi-Te Kao, 高啟德
• Other Authors: Jia-Sheng Heh
• Format: Others
• Language: zh-TW
• Published: 2011
• Online Access: http://ndltd.ncl.edu.tw/handle/32545339068982942808

碩士 === 中原大學 === 資訊工程研究所 === 99 === Researchers had found that college students in Taiwan who learn English as a Foreign Language (EFL) experienced difficulties in English reading comprehension. Several kinds of knowledge is needed in reading comprehension including vocabulary (word sense) and structural (sentence parsing) knowledge. These knowledges are implicitly learned by comprehensive students, but are sometimes difficult for other students to learn. Thus, an explicit representation of these knowledge is needed. Previous works on sentence parsing are doing full parsing through prescribed grammars like context free grammar (CFG), however, it is difficult for students to apply these grammars to fully parse English sentences. There is a different approach of parsing, called partial parsing, which can be used to easily and robustly parse sentence into chunks (noun phrase, verb phrase, preposition phrase, simple clause, etc). On the other hand, the knowledge of word sense can be represented through construction grammar by explicitly labeling the case roles of words that are used together in chunks. The purpose of this paper is to develop an interactive tool, based on partial parsing and construction grammar, to capture and explicitly represent the knowledge and rules needed for students to parse and understand English sentences. The interactive tool will be provided as an auxiliary learning tool for students in reading comprehension. The tool includes: a student editable rule-based parser, which is based on partial parsing, with both sequential and global approaches; an interactive rule-defining interface for students to define their own rules which can be organized and applied later; a knowledge representation model based on construction grammar to represent the semantic knowledge of words in the sentences; a student editable lexicon bank for recording their own lexical knowledge. The tool is implemented and applied as an auxiliary tool for studying and answering the questions of reading comprehension test of a college entrance examination of English in Taiwan. A set of parsing rules and semantic knowledge models are collected and presented. A difficulty index of sentences such as number of chunks, number of coreferences, and depth of stack is also presented.