資訊管理學報
羅靖華;黃文楨;陳欣蓉;
頁: 1-24
日期: 2011/01
摘要: 在學習語言的過程中，除了學習單字與單辭之外，能夠有完整無錯的句子更是重要。但是目前所有字典辭典系統的例句，依然侷限在編寫者依照查詢的單字或單辭來編寫的。這種作法比較傳統而且通常不能切入使用者學習上的需要。本系統的目的就是要解決這樣的語言學習困難，以目前較少人開發的德文學習這個區塊，使用「以句找句」的方式，輸入中文詞句就會找出結構最相近的德文詞句，對於語言的學習很有幫助。本研究突破傳統詞彙比對的作法，以樹狀結構表示句型結構，提供有序與高擴展性的句型結構框架，建立以句構與辭性為基礎的索引機制。由全句的句子組成結構與其中各單詞的辭性進行快速索引，並以本體論(Ontology)結合主題地圖(Topic Maps)概念建構不同情境的應用領域路徑，克服迥異情境下辭彙取用問題，實作系統後透過問卷調查得知，服務機制獲得使用者認同，反映出本研究提出的創新方式對於語言學習將確實有所助益。
關鍵字: 機器翻譯;本體論;主題地圖;
Abstract: In a language learning process, it is important to express complete and correct sentences in addition to learning vocabulary and idioms. However, the sample sentences in current dictionaries are based on user's querying specific vocabulary, and do not usually meet user's wider learning requirements. The purpose of the proposed system is to solve such a language learning difficulty by using a ”sentence to sentence” mechanism for Chinese to German sentence query and translation. It allows inputting a Chinese sentence and the system then displays several German sentences with similar structure/meaning by descending similarity. The proposed system, based on a sentence parsing tree structure, provides an ordered and highly extensible sentence structure framework to construct a semantic and grammar-based indexing utility. Fast indexing is processed by parsing a sentence structure and extracting the syntactical function of each word. It is supported by Ontology and Topic Maps to construct various application domain paths for vocabulary extraction while in different scenarios. Questionnaire survey was informed that this innovative study for language learning would be helpful.
Keywords: Machine Translation;Ontology;Topic Maps;
瀏覽次數: 13172 下載次數: 122
引用 導入Endnote
頁: 1-24
日期: 2011/01
摘要: 在學習語言的過程中，除了學習單字與單辭之外，能夠有完整無錯的句子更是重要。但是目前所有字典辭典系統的例句，依然侷限在編寫者依照查詢的單字或單辭來編寫的。這種作法比較傳統而且通常不能切入使用者學習上的需要。本系統的目的就是要解決這樣的語言學習困難，以目前較少人開發的德文學習這個區塊，使用「以句找句」的方式，輸入中文詞句就會找出結構最相近的德文詞句，對於語言的學習很有幫助。本研究突破傳統詞彙比對的作法，以樹狀結構表示句型結構，提供有序與高擴展性的句型結構框架，建立以句構與辭性為基礎的索引機制。由全句的句子組成結構與其中各單詞的辭性進行快速索引，並以本體論(Ontology)結合主題地圖(Topic Maps)概念建構不同情境的應用領域路徑，克服迥異情境下辭彙取用問題，實作系統後透過問卷調查得知，服務機制獲得使用者認同，反映出本研究提出的創新方式對於語言學習將確實有所助益。
關鍵字: 機器翻譯;本體論;主題地圖;
Construction of Semantic and Sentence Patterns Retrieval Service System
Abstract: In a language learning process, it is important to express complete and correct sentences in addition to learning vocabulary and idioms. However, the sample sentences in current dictionaries are based on user's querying specific vocabulary, and do not usually meet user's wider learning requirements. The purpose of the proposed system is to solve such a language learning difficulty by using a ”sentence to sentence” mechanism for Chinese to German sentence query and translation. It allows inputting a Chinese sentence and the system then displays several German sentences with similar structure/meaning by descending similarity. The proposed system, based on a sentence parsing tree structure, provides an ordered and highly extensible sentence structure framework to construct a semantic and grammar-based indexing utility. Fast indexing is processed by parsing a sentence structure and extracting the syntactical function of each word. It is supported by Ontology and Topic Maps to construct various application domain paths for vocabulary extraction while in different scenarios. Questionnaire survey was informed that this innovative study for language learning would be helpful.
Keywords: Machine Translation;Ontology;Topic Maps;
瀏覽次數: 13172 下載次數: 122
引用 導入Endnote