English  |  正體中文  |  简体中文  |  Items with full text/Total items : 90453/105672 (86%)
Visitors : 12011980      Online Users : 616
RC Version 6.0 © Powered By DSPACE, MIT. Enhanced by NTU Library IR team.
Scope Tips:
  • please add "double quotation mark" for query phrases to get precise results
  • please goto advance search for comprehansive author search
  • Adv. Search
    HomeLoginUploadHelpAboutAdminister Goto mobile version
    ASIA unversity > 資訊學院 > 會議論文 >  Item 310904400/5564


    Please use this identifier to cite or link to this item: http://asiair.asia.edu.tw/ir/handle/310904400/5564


    Title: 以先進科技促進社會互動之高階英語?位學習與自動評?系統之研究
    Authors: 張俊盛;?顯親
    Contributors: 清華大學資訊工程系
    Date: 2007-12-07
    Issue Date: 2009-12-15
    Publisher: 主辦單位:國科會科學教育處;承辦單位:亞洲大學資訊學院
    Abstract: 本研究旨在?用先進的資訊與網?技術,研發適合台灣的教學與檢測方式,提供?多的自動化資源,輔助教師準備教材、出題、閱卷,並讓學生能夠透過與電腦系統與同儕互動,強化自主學習的效果。第一?我們進??多項研究:包括學術英文單字語聽?教材的研發,學生語?庫的建?語相關研究、辭彙搭配學習工具的語意組織等。
    首先,在學術英文單字學習方面,我們研究將單字學習融入閱?活動,並實驗探??用學術英文詞彙表(Academic Word List)與語?庫?引的工具,探?直接學術英文詞彙教學(explicit academic vocabulary instruction)的效果。研究結果顯示,學生單字深?(depth)能?有顯著的進步。其次,我們也研究AWL在應用語言學文獻中的分布情形與辭彙教學上的意涵。另外在聽?教學上,我們提出一個以三個因素構成之評?聽?教材難??的公式。藉著此公式,我們架構?一聽?網站以訓?學生的聽?,同時也藉著學生們在網站上的活動記?瞭解公式與教材的效用。研究發現線上聽?教材對學生聽?確有助?。另外公式似乎和學生們的難??認知有所出入,尚須進一步修正。
    第二項工作是蒐集學生語?庫的蒐集,以配合研究如何指導、評估學生作文。語?庫的蒐集範圍包括台大、台中師大、東吳大學主修英文學生的作文,總?將超過百萬字。預計配合自然語言處?與機器學習技術,發展寫作錯誤偵測與回饋、寫作評分的自動化機制。我們初步英語?庫為基礎,開發一套網?版軟體,目前已經有效偵測出幾?的學生作文錯誤,包括拼寫、搭配、主動詞一致等錯誤。
    第三項工作是持續改?CANDLE計畫發展出?的一些工具,包括TotalRecall和TANGO。今??我們研究透過語意分析,組織 TANGO 所呈現的??很多的搭配詞。我們發現?用WordNet 的辭彙相似?計算,可以將特定名詞的V-N搭配詞加以合?的分?,由助於?有效?呈現搭配詞,讓學生?於使用。 To help language teachers, university administrators, and academic advisors easy the burden of teaching, testing, and revision, we propose to exploit existing language corpora andstate-of-the-art technologies in Natural Language Processing and Internet communication to assist the teaching and testing of academic reading, writing and speaking. The preliminary results of the first year into the three years duration of this project include academic vocabulary building through explicit learning, listening difficulty measure, learner corpus and related study, semantic organization of collocation reference tool.
    First, in the area of academic vocabulary learning, we investigated the effects of explicit academic vocabulary instruction using AWL and concordance of the BNC. The results indicate that significantly enhanced the depth of vocabulary. We also explored the AWL distribution of AWL in literature of applied linguistics study and its implication for vocabulary teaching. And in the area of listening comprehension, we examined to what extent online listening materials can benefit the learners in their listening, and proposed a formula to grade listening texts. We found that the online listening materials were useful in enhancing the learners’ listening ability and listening difficulty are more complex issues than we have assumed.
    The second part of our effort involves learner corpus. We have compiled learner corpora with corpus size exceeding more than one million words to put to use in studying computer-assisted writing and automatic grading of writing. We now have a expanding corpus containing writings by students in National Taiwan University, National Taiwan Normal University, National Tsing Hua University, and National Taichung University, and Soochow University. In addition, the learner corpus has been augmented with students writings arising from online intercultural dialogues between EFL students and pre-service teachers.
    We also developed a computer assisted writing tool based on reference corpora. Using very simple n-gram lookup, the online tool is capable of identifying a number of errors including spelling errors, miscollocations, and agreement errors.
    Finally, we are devoted to further improve the tools developed in the CANDLE Project. We use WordNet-based similarity measures to organize the extremely numerous collocates that these tools can return into semantically relevant clusters, focusing on V-N pairs and clustering over the verbs. The method based on unsupervised learning indeed discovers verb clusters for collocations with reasonable performance.
    Keywords: digital language learning, computer assisted language learning, computer assisted testing, text corpus, computational scaffolding, formative evaluation, summative evaluation, Corpora And NLP for Digital Learning of English (CANDLE)
    Relation: 國科會科教處95年度資訊教育學門專題研究計畫成果討論會
    Appears in Collections:[資訊學院] 會議論文

    Files in This Item:

    There are no files associated with this item.



    All items in ASIAIR are protected by copyright, with all rights reserved.


    DSpace Software Copyright © 2002-2004  MIT &  Hewlett-Packard  /   Enhanced by   NTU Library IR team Copyright ©   - Feedback