全球化地球村已經來臨了，而言語的溝通在這中間扮演非常重要的角色，英語的重要性更是不可言喻。英語學習者若只是一直閱讀文章，而不理解關鍵字彙、片語對閱讀文章整體的重要性，將影響對文章內容的理解，更導致測驗答題正確性降低。
本研究提出並開發一套運用網際網路線上字典為基礎的英文測驗自動出題系統，結合剖析器的語法分析英文句子結構和字彙的詞性標註，再加上本研究提出的三個英文試題自動出題模組，自動產生英文測試題目，不僅有助於老師節省大量時間和精力去命名出題，更有助於學生複習時的成效，增強學生的英文能力。透過實作驗證，更確認本系統產生的測驗題目具有可行性和可用性。

Network Dictionary Based Automatic English Test Generating System
Hung Wen Chu
Tzone I Wang
Department of Engineering Science
National Cheng Kung University

SUMMARY

Global village has been already shaped, in which verbal communication plays a very important role, and the importance of English is also ineffable. For English learners, if they only read articles without understanding the sacred words and phrases on the importance of the article, the content of the articles will be misjudged which results in reduced correctness on answers to the test for the articles. This research proposed and developed an automatic English test items generation system using a set of Internet-based online dictionaries. The system uses a parser to analyze the structure of English sentences extracted from interested articles found in the Internet and to find the POS tags of vocabulary for the automatic generation of test items by the three modules. This system not only helps teachers saving a lot of time and effort to compile tests for a topic, but also, by generating test items automatically, helps students effectively review the topic and enhance students' English proficiency. Reviewed by some experts, the test items generated by the system had been proved to be feasible and effective for different kinds of subjects.

Keywords: Automatic Test Item Generation System, E-learning, English Learning

INTRODUCTION

To meet the needs of learners of English learning, the study designs an automatic English test generation system that combines of network resources. After English learners learn to read the article, the system can been automatically generated tests. This not only helps learners understand the content of the article, but also take refresher deepen learning.

LITERATURE
e-Learning research
e-Learning is the learner through some digital media, such as Internet, computers, mobile devices, satellite radio, etc., to carry out learning activities. Learners through these digital media, connect learning platform, and get on the platform multimedia text, pictures, video, etc. without being limited to the time and distance to achieve ubiquitous learning.

Computer-assisted item generation
With the rapid development of information technology, the traditional paper questions quiz has been slowly declining, and been replaced by a computer-assisted item generation. With the powerful computing power, this system can reduce a lot of manpower and time consuming, and can quickly output quantities and quality of these items with a certain. And computer-assisted item generation also supports adaptive learning system at different stages of the learner, and provides different levels of items to help learners learn and grow gradually progressive.

Natural language processing research
Natural language processing is a very important popular area in language studies and artificial intelligence research. As the name suggests, natural language processing is to translate human language, through natural language understanding analysis, conversion, and generates the language which machine can understand.

MATERIALS AND METHODS
This study suggests that verbs are the most important keywords in the English sentence. If the verb's synonymous and antonym are used to generate tests ,it not only learn how to use the verb in the sentence, but also learn more of replications. English phrases in the English sentence can increase the depth and splendor of the English sentence. The relative pronoun in the English sentence can play go-between functions. It also combines with a series of other sentences, and increases the integrity of the English sentence. Therefore, the phrase generating tests and relative pronoun generating tests are also the emphasis of automatic generating tests.
Teachers through the system as long as the user interface can import textbook articles or articles in English easily by the way files. And then English sentence parser module can do a preliminary analysis of punctuation and sentence, and the results are stored in a exam database. Verb's synonymous and antonym tests module can automatically select the appropriate verb from the sentence, and generate the verb options and topics. Phrase tests module can automatically select the appropriate phrase from the sentence, and generate the phrase options and topics. Relative pronoun tests module can automatically select the appropriate relative pronoun from the sentence, and generate the relative pronoun options and topics. Finally, teachers press the show test button in generating test page of the system and the completed tests will been showed in the text window.
RESULTS AND DISCUSSION

This study randomly selected a currently high school text in high school English textbook and used it to do the actual testing. There are 37 English sentences which been imported by manual. And then these tests through the system's three processing modules were generated a total output of 29 tests available for learners. If the ratio is calculated according to the all tests and then there are the tests about the rate of close to 80% faster. The tests through automatic generating test system have been confirmed by professional English teachers and been confirmed automatic generating tests correctly.

CONCLUSION

The system interface is simple, easy and convenient operation, without specially operational training. Just a few steps, you can easily complete a simple generating tests automatically. You can also easily export the tests easily, and turn into a paper test. This study use of Internet resources readily available and generate tests at low cost. It not only help teachers focus teaching the topic without time-consuming, but also help students review and strengthen the English learning. Tests which been generated by the automatic generating test system have been proven the feasibility and availability.

[1] Raab, R. T., Ellis, W. W., & Abdon, B. R. (2002). Multisectoral partnerships in
e-learning A potential force for improved human capital development in the
Asia Pacific. Internet and Higher Education , 217–229.
[2] Chen, C., & Chung, C. (2008). Personalized mobile English vocabulary learning
system based on item response theory and learning memory cycle. Computers &
Education, 51(2):624-645.
[3] Wang, k.,Huang, Y., Jeng, Y., & Wang, T. (2008). A blog-based dynamic learning
map. Computers & Education, 51(1):262-278.
[4] Wang, Y., Tseng, M., & Liao, H. (2009). Data mining for adaptive learning
sequence in English language instruction. Expert systems with
applications:36(4),7681-7686.
[5] Hsu, M. (2008). A personalized English learning recommender system for ESL students.Expert systems with applications, 34(1):683-688.
[6] 傅茹舷(2009)，利用記憶週期更新與內文關係增進英文字彙的學習，國立成功大學工程科學所碩士論文。
[7] 施弼耀(2006)，線上動態英語閱讀學習平台建置與使用接受度評估，屏
東教育大學學報 Vol.24 pp.521-554.。
[8] Deane, P. & Sheehan, K. (2003). Automatic item generation via frame semantics.
Education Testing Service: http://www.ets.org/research/dload/ncme03-
deane.pdf .
[9] L. A. H. Ruslan Mitkov (2003). Computer-Aided Generation of Multiple-Choice
Tests,Proceedings of the HLT-NAACL 03 workshop on Building educational
applications using natural language processing -Volume 2, 17-22.
[10] 王俊弘，劉昭麟，高照明(2004)，利用自然語言處理技術自動產生英文克漏詞試題之研究。
[11] G. A. F. M. E. Jonathan C. Brown(2005).Automatic Question Generation for Vocabulary Assessment,HLT '05Proceedings of the conference on Human Language Technology and Empirical Methods in Natural Language Processing, 819-826.
[12] 陳佳吟、柯明憲、吳紫葦、張俊盛(2005)，FAST：電腦輔助英文文法出題系統，第十七屆自然語言與語音處理研討會。
[13] 楊媛茜、楊捷扉、張嘉銘、張俊盛(2005)，電腦輔助閱讀測驗自動出題，第十七屆自然語言與語音處理研討會。
[14] 史馥銘(2005)。利用自然語言處理技術之英文試卷自動出題，電腦與通
訊，第 112期，pp.70-74。
[15] Chen, K., & Chen, H. (1993). A probabilistic chunker. Paper presented at the Proceedings of R.O.C. Computational Linguistics Conference.
[16] Hindle, D. (1983). User manual for Fidditch, a deterministic parser.Naval
Research Laboratory Technical Memorandum,7590-7142.
[17] Goeman, H. (2001). On parsing and condensing substrings of LR languages in linear time. Theoretical Computer Science,267(1-2):746.
[18] Shishibori, M., Sangkon Lee, S., Oono, M., & Aoe, J. (2002). Improvement of
the LR parsing table and its application to grammatical error
correction .Information Sciences, 148(1-4):11-26.
[19] The Stanford Parser: A statistical parser (version 2.0.4),
http://www-nlp.stanford.edu/software/lex-parser.shtml [Last visited on
November 2012].
[20] Klein, Dan. and Manning, Christopher. (2003) Accurate Unlexicalized Parsing.
Proceedings of the 41st Meeting of the Association for Computational
Linguistics, 423-430.
[21] WordNet, http://wordnet.princeton.edu/.
[22] DICT.TW線上字典, http:/ dict.tw/index.pl/.
[23] 譯典通線上字典,http://dict.dreye.com/ews/get--01--.html/.