节点文献
面向E-learning课件的个性化检索系统的研究与实现
The Design and Implementation of Personalized Indexing System Based on E-learning Courseware
【作者】 陈悦;
【导师】 申瑞民;
【作者基本信息】 上海交通大学 , 计算机软件与理论, 2008, 硕士
【摘要】 随着网络教育的蓬勃发展,与之相适应的资源建设也日益受到人们的重视。众所周知,网络教育有三大基础:硬件、软件和网络教育资源。Internet技术正在向宽带、高速、多媒体化方向的快速发展,硬件、软件方面的优势将逐渐弱化,教育资源的优势必将越来越重要,最终将成为决定网络教育成败的关键所在,而网络课程是其中非常重要的组成部分。在远程教育环境中,网络课程因不受时间空间的限制,使用灵活,也因其录制于真实课堂而获得的可视性、现场感,使得学生的学习时间更加灵活,学习效率更高,因此成为了学生的主要学习资源。然而,网络课程资源的不断增加,给资源检索带来了很多困难。每天都有大量的新资源被增加到学习系统中,而且学习者一般只对符合自己需要的课程感兴趣。所以,为网络课程资源建立索引,向学习者推荐他们感兴趣的学习资源,成为当前网络课程的急切需求。本文主要工作包括:(1).在多媒体课件制作部分中,我们提出了多媒体课件录制的实时同步融合技术,将老师视音频和操作三路信息一并作为脚本同步写入WMV文件。同时,我们采用微软的OLE技术转换PPT至Web格式。这两种技术的有机结合使得生成的多媒体课件具有清晰度高、占用资源小、同步效果好等优点。同时,利用WMSDK类库实现对WMV文件的自由剪切和合并。(2).在教学资源个性化检索部分中,我们提出基于学习者兴趣特征向量的课件检索方案。根据学习者点播课件后给出的打分,确定学习者的兴趣特征向量,形成当前学习者的推荐课件列表。同时,通过学习者之间相互推荐学习资源,确定每个学习者各自的学习好友、潜在好友、陌生人和未知人四个列表,在具有相似学习兴趣的好友之间,相互推荐学习资源,形成个性化的学习资源检索。这样,我们的检索具有更高的正确率。
【Abstract】 With the development of network education, people put more focus on the corresponding network resource construction. As we all know, network education consists of three sections: hardware, software and education resource. Internet technique is quickly developing in a board band, high speed and multimedia orientation, and this weakens hardware and software’s advantage, and the education resource becomes more and more important, which will finally determine whether the network education can succeed. The network coursewares are the main factor of network education. In a distance-learning environment, coursewares can make students learn more easily with an unlimited space and time, which becomes students’main learning resource. With the increase of coursewares, the indexing work for these courses becomes more difficult.Everyday large amount of resource are added into research lab, and people always take more care about the course they interest in. As a result, there is an urgent need to make an index for these coursewares and recommend some suitable course for online learners.Main issues of this article include:(1). In multimedia courseware generating subsystem, we proposed a multi-flow synchronized fusion of the whole-scene data technology, which can merage teachers’audio, vedio, and action into a WMV file as scripts. On the other hand, we achieve a web format conversion of PPT. Cmbining these two technology, we can gain a real-time and, efficient generation of high-quality streaming media courseware. At the same time, we finish the cutting and merging work for WMV file using WMSDK library.(2). In personalized multimedia indexing subsystem, we proposed a multimedia courseware indexing technology based on learners’interest feature vector. After users log in the distance learning system, they can watch and rate courseware and the user interest feature vectors are formed according to our rating value. At the same time, we can construct our learning friends, potential friends, stranger and unknown indexing with the rating value and user’s interest feature vector. These indexing can provide suitable learning resource and enhance users’learning effectiveness. On the other hand, we can merge it into two other obtained indexing functions: title based and contented based, which can highly improve the indexing function for multimedia courseware.
【Key words】 multimedia; personalized indexing; courseware weblization; interest feature vector; peer to peer;
- 【网络出版投稿人】 上海交通大学 【网络出版年期】2008年 06期
- 【分类号】TP391.3
- 【下载频次】158