摘要
随着网络的不断发展和普及，网上教学已经成为培养人才、促进科研和教育事业发展的重要途径。然而要使网络和信息技术真正为教学服务，实现教学过程和教学资源的优化，就必须有丰富的教学资源支持，发展网上教学的先决条件是构建完整、充实的网上教学资源体系。
目前的教学资源库系统主要提供教学资源上传、教学资源查询和教学资源下载等功能。主要就是以“物”为主，未能体现“以人为本”的思想，特别是现有的教学资源库不具备个性化和智能化等特性，从而普遍存在海量教学资源与用户个性化需求之间的矛盾。由于这个矛盾的存在，一方面造成了教学资源的巨大浪费，另一方面，为用户查找和使用所需要的教学资源造成了巨大的因难。这些问题都阻碍了教学资源在教学中作用的发挥，造成了极大的教学资源浪费。
本文的主要工作如下：(1)针对目前教学资源服务中普遍存在的海量教学资源与用户个性化需求之间的矛盾，论文在网络教学资源库系统中建立了一个教学资源个性化服务模型(Individualized
Service Model of Treaching Resource，以下简称ISMTR)，该模型把基于网络的个性化信息服务引入教学资源体系，尊重用户中存在的差异，学习和跟踪用户的个性化兴趣，并根据用户的个性化兴趣特征对教学资源进行过滤，实现少而精的教学资源个性化检索服务和推荐服务；(2)根据该ISMTR模型，在网络教学资源库中设计了个性化教学资源过滤、个性化检索和个性化推荐等个性化教学资源服务模块；(3)对实现教学资源个性化服务的教学资源的表示、用户兴趣特征提取、用户模型的构建、教学资源过滤技术等关键技术及实现方法进行研究与实验；(4)建立了基于ISMTR模型的ITRDS实验系统及其对照实验系统，并进行了相应实验。
本文的主要工作特点有：(1)在用户个性化特征信息的保存方面，采用用户个性化特征库来记录和跟踪用户对教学资源的个性化访问特征，以便进行个性化服务。(2)在用户模型的构建方面，由系统对用户的个性化资源访问进行特征提取，构成个体用户模型，并对兴趣相似的用户进行聚类构成小组用户模型，最后把这两个用户模型进行综合，形成综合用户兴趣向量以构建综合用户模型。(3)在过滤技术的选择方面，采用基于个体用户模型的个体过滤、基于小组用户模型的协同过滤和基于综合用户模型的综合过滤这三种过滤技术对教学资源进行过滤。其中，综合用户模型不仅反映该用户的个性化兴趣特征，它还包含该用户所在类的兴趣特征，所以基于综合用户模型的教学资源综合过滤既有个体过滤的查准率高的优点，又有协同过滤的查全率高的优点。
论文的主要研究意义有：(1)提出个性化教学资源服务的实现方法，以此解决海量教学资源与用户个性化需求之间的矛盾，有利于提高网络教学资源库系统的个性化和智能化；(2)提出一种兼具个体过滤技术和协同过滤技术优点的综合过滤技术，对其它基于网络的个性化信息服务如电子商务有借鉴意义；(3)提出采用实际兴趣指标和综合兴趣指标相结合来描述用户兴趣，并采用用户个性化特征库来记录和跟踪用户对教学资源个性化访问，给个性化教学资源服务提供一种新思路。
With the development and widespread of network, E-teaching has become one of the most important ways of educating and researching. In order to make the network and IT actually helpful to education, it's essential to optimize the teaching process and resources. The precondition to develop E-teaching is to establish an integrative system of teaching resources.
The present system of teaching resources mainly serves to supply developed courseware and teaching plans for users to transfer and download. It is not "user-centered", but "material-centered". Because the present teaching database isn't individualized or intelligentized, it is not convenient to manage and finding the wanted resources. As there is a wide variety of styles, expectations, paces of different learners, there is commonly a controversy between the sea of teaching resources and the individualized demands. Therefore, it brings about much waste of teaching resources. And on the other hand, it is very difficult for users to consult and find what they want. These problems prevent the teaching resources from working effectively in teaching.
Aimed at the contradiction between individualized requirements and the masses of teaching resources, this paper designed an individualized service model of teaching resources, which employs the individualized information service on the Net into teaching. It defers to the differences of users, studies their habits and interest. By filtering the teaching resources, the precision researches and recommendation can be made. On the basis of ISMTR, this paper designed individualized modules for filtering, searches and recommendation. Moreover, the paper studied the individualized denotation of teaching resources, the extraction of users' interest, the establishment of users' models and some important technologies such as filtering, it also developed an experiment (ITRDS), and used the experiment to verify the realization method.
The paper has several specialties. On the information store of individualized characters, the paper made use of individualized character database to record the users' individualized accessing. The system can extract the characters of the users to get the eigenvector of every individual and cluster the similar users to get the eigenvector of groups. Then these two eigenvectors will be integrated to form the integration eigenvector. The model realizes information filtering by individual filtering, collaborative filtering and integration filtering. The group users model contains the interest of individuals and the interest of groups. So the integration eigenvector will certainly result in high precision of individual filtering and high recall of collaborative filtering.
There are a few significances. This paper put forwards the realization of the individualized service model of teaching resources to solve the contradiction between individualized requirements and the masses of teaching resources, and it's helpful to make the teaching resources database individualized and intelligentized. The integration filtering can be used in reference in other individualized information services such as E-business. It also describes the users' interest by the real interest parameter and the integrated interest parameter, and made use of individualized character database to record the users' individualized accessing.
引文
[1] 刘风新．基于Java技术实现交互式个性化的远程教学系统[J]，北京化工大学学报，2003．2
[2] 丁琳．数据挖掘在远程教育个性化服务中的应用[J]，电化教育研究，2002．9
[3] 邢东山．基于WEB使用挖掘技术的个性化教育网站构筑[J]，计算机应用．2003．1
[4] 王继成等．基于Internet的教学资源发现技术与实现[J]．计算机研究与发展．1999．11
[5] 沈军．网络教学中个性化策略研究[J]，计算机研究与发展2003 40(4)
[6] 田萱等．实现Web页面的智能个性化检索[J]．计算机工程与应用．2003．01
[7] 宋玲等．Internet个性化智能信息检索的分析与研究[J]．情报学报．2002．1
[8] 丁琳等．数据挖掘在远程教育个性化服务中的应用[J]．电化教育研究．2002．09
[9] 李宝敏．从知识管理的角度看远程教育中个性化资源库的建设[J]，中国远程教育，2003 192(3)30-32
[10] 寇兴权，刘兴环．远程教育资源的描述、组织和管理系统设计[J]，中国远程教育，2002．9(4)35～37
[11] 余武．信息化教学资源的开发和建设[J]，中国电化教育，2001．7，15～18
[12] 曾春，邢春晓，周立柱．基于内容过滤的个性化搜索算法[J]，软件学报，2003．14(5)999～1004
[13] Nils J Nilsson著，郑扣根译．人工智能[M]，北京：机械工业出版社．2001年8月．
[14] 徐小琳，阙喜．信息过滤技术和个性化信息服务[J]，计算机工程与应用，2003 9 182～184
[15] 李勇，徐振宁，张维明．INTERNET个性化信息服务研究综述[J]，计算机工程与应用．2002 19 183～188
[16] 范雪敏．个性化电子商务网站的研究与实现[J]，计算机应用．2002，6
[17] M de Kroon et al. Improving learning accuracy in information filtering. http://cs.cmu.edu
[18] 王向星．智能化个性服务在Web上的应用研究[J]．计算机与现代化．2003．3
[19] 李煊．基于关联规则挖掘的个性化智能推荐服务[J]，计算机工程与应用．2002．12
[20] 王实等．基于分类方法的Web站点实时个性化推荐[J]．计算机学报．2002．08
[21] 肖小军等．一个基于因特网的个性化信息服务系统的设计和实现[J]．计算机工程与科学．2002．01
[22] 教育部．现代教育技术规范[M]．2001
[23] 肖晓军．一个基于因特网的个性化信息服务系统的设计和实现[J]．2002(24)，1
[24] 赵涓涓，陈俊杰．信息过滤中用户个性化模式的构建[J]，太原理工大学学报，2003 34(3)336～339
[25] Olivia P R著，朱扬勇等译，数据挖掘实践[M]．机械工业出版社，2003．09
[26] Dumais S T, at al. Inductive learning algorithms and representations for text categorization. Proceedings of the International Conference on Information and Knowledge Management. New York: ACM Press, 1998. 148～155.
[27] Letsche T A, at al. Larger scale information retrieve with latent semantic indexing [J]. Information Sciences, 1997, 03
[28] Moldova D I, at al. Using Word net and lexical operators to improve Internet search [J]. Internet computer. 2000.02
[29] Mehmed K著．闪四清等译，数据挖掘——概念、模型、方法和算法[M]．清华大学出版社．2003．08
[30] Pazzani M, Billsus D. Learning and revising user profiles: The identification of interesting Web site. Web Learning. 1997, 27(2)313～331
[31] 申瑞民．个性化数字服务模型，微电子学与计算机[J]，2001．1，14-18
[32] David A. hull. The TREC-6 Filtering Track Description and Analysis. http://trec.nist.gov/pubs/
trec6/papers/filter.track.ps.gz
[33] 李煊，基于关联规则挖掘的个性化智能推荐服务[J]，计算机工程与应用．2002．12
[34] Tu H C, at al. An architecture and category knowledge for intelligent information retrieval agents [J]. Decision support system. 2000.06
[35] 张国印．陈先，皮鹏．基于词频统计的个性化信息过滤技术[J]，哈尔滨工程大学学报．2003 14(1)63～68
[36] Jiawei Han, Micheline Kamber，著，范明译，数据挖掘技术[M]．北京：机械工业出版社，2001年8月。
[37] 赵亮．个性化推荐算法设计[J]，计算机研究与发展，2002，39(8)
[38] Yah T W, at al. Index structure for information filtering under the vector space model[R]. Technical Report STAN CS 93. Stanford University. 1993