《毕业设计外文文献—在线课程视频观看分析.docx》由会员分享,可在线阅读,更多相关《毕业设计外文文献—在线课程视频观看分析.docx(14页珍藏版)》请在taowenge.com淘文阁网|工程机械CAD图纸|机械工程制图|CAD装配图下载|SolidWorks_CaTia_CAD_UG_PROE_设计图分享下载上搜索。
1、附录A外文翻译一原文局部出处: s:/ieeexplore.ieee.org/document/7973527Analysis of video views in online coursesEsztelecki PeterUniversity of Szeged, Faculty of Science and Informatics epeterinf.u-szeged.huHavasi FerencUniversity of Szeged, Faculty of Science and Informatics havasiinf.u-szeged.huAbstract - Nowadays
2、, in our fast-paced world there are countless MOOC courses in the Internet with various topics that have been designed to broaden our knowledge. One of the most powerful tools fbr effective learning are online videos. Many case studies have been carried out in order to specify the qualities of a goo
3、d educational video. Philip J. Guo, Juho Kim and Rob Rubin published an article (How Video Production Affects Student Engagement - An Empirical Study of MOOC Videos) on their experience during edX courses. They analyzed the optimal length of a video and determined which videos were preferred by the
4、students.Our research group from the University of Szeged has developed a logging system which records student activities in online courses to analyze learning styles and behaviour when watching a video. In our research, we did a literature review about on-line videos and introduced the system to lo
5、g student activities, focusing on video views.Keywords: on-line videos, video views, MOOC courses, electronic cumculumI. INTRODUCTIONE-leaming (Electronic Learning), in our times, is perceived as a new technology in the field of education, which makes possible sharing and reaching learning content,
6、practicing and measuring acquired knowledge, and finally communicating between members of courses. Education in case of E-leaming is computer supported which allows individual learning pace cancelling traditional school timetables fbr the benefit 附录B外文翻译一译文局部在线课程视频观看分析Esztelecki P6ter塞格德大学,科学与信息学学院e
7、peterinf. u-szeged. huHavas i Ferenc塞格德大学,科学与信息学学院havasiinf. u-szeged. hu摘要-在当今快节奏的世界里,网络上有无数的MOOC课程,其主题五花八门, 旨在拓宽我们的知识面。有效学习最强大的工具之一是在线视频。为了明确一个 良好的教育录像的质量,已经进行了许多案例研究。Philip J. Guo, Juho Kim和 Rob Rubin发表了一篇关于他们在edX课程中的经历的文章(视频制作如何影响 学生参与一一MOOC视频的实证研究)。他们分析了一段视频的最正确长度,并确 定了学生们更喜欢哪些视频。我们来自深圳大学的研究小组开
8、发了一个记录学生在线课程活动的日志系统,用 来分析学生在观看视频时的学习风格和行为。在我们的研究中,我们对网络视频 进行了文献综述,并引入了该系统来记录学生的活动,重点是视频观看。关键词:在线视频,视频观看,MOOC课程,电子课程I.介绍在我们这个时代,E-learning(电子学习)被认为是教育领域的一项新技术,它 使得学习内容的提供和到达,学习知识的实践和测量,以及课程成员之间的交流 成为可能。在电子学习的情况下,教育是计算机支持的,它允许个人学习速度取 消传统的学校时间表,以获取知识的最方便的时间。要开展网络学习,不仅要大 力推广终身学习,更要大力推进素质教育,倡导国际教育;尽管最重要的
9、因素是 信息技术的加速进步。另一种主张通过移动设备学习的观点也得到了支持,即所谓的移动学习。它 强调知识不仅依赖于时间,而且可以获得,无论地点.这样的技术进步可以简 化终身学习的实现,因为高级学习将与全职或兼职工作并列进行。越来越多的大 学开设网络课程,甚至渗透到中小学,教师将实验网络课程纳入课程。我们可以区分虚拟方法和混合方法。在虚拟学习环境(VLE)中,通过提供信 息和参与论坛讨论,实现了学习材料的共享和与虚拟方法的交流。因此,学习者 不与导师见面,也不与他人见面,这在成人教育中不是问题,但对于K-12年龄 段的学生来说并不方便,他们不能忽视监督,也不能为了相互学习而与同龄人见 面。在后一
10、种情况下,一种混合的方法,即混合学习,将更有益,因为它将意味 着在一个公认的课程范围内的古典学校课程,然而,一些要素仍然可以通过在线 方法和个别时间表获得,而不管地点如何。混合学习使学习者能够加深他们的知 识,这些知识可以通过测试来衡量,而且似乎更持久n.网上课程的视听内容视听内容以听觉和视觉能力为目标,以触发信息获取。这样的工具和学习材 料出现在第一次世界大战后作为实物课程,虽然在那个时候幻灯片投影机已经被 使用,并配有录音磁带。在60年代和70年代,电视的群众普及到达了顶峰,正 是在那个时候,炒作周期代表着更高的期望。在美国进行的大多数研究都试图证 明通过看电视进行的教育比传统方法更有效,
11、但很快就发现并不存在一种单一的 最有效的系统。后来,随着个人电脑在家庭中出现,互联网覆盖全球,90年代 出现了新的机遇,观看和提供视频材料帮助建立了第一批在线课程。第一代数字学习材料只不过是扫描传统教科书和数字化教育视频,而第二代 工具完全使用数字媒体和为计算机使用优化的内容组合在一起。第三代数字教育 材料的标准如下:互动性,包括广泛的电影和音频记录,使用链接加强导航,方法 指导,所有这些都创造了一种独立的学习材料。第四代强调互动性和协作,其产 品并非源自单一作者,而是可以来自各种多元文化社区。在YouTube等在线视 频提供网站推出后,这一领域的真正突破是显而易见的,因为从那以后,教育视 频
12、变得很容易访问。例如,可汗学院的视频提供自2006年以来在YouTube上被 浏览了 6.63亿次。MOOC (Massive Open Online Course,大规模公开在线课程)在 这一时期广为人知,它包括无限参与和免费访问的网络课程。我们可以区分四种 类型的教育视频:(1)课堂教学,基本上是学校课程或大学讲座的记录(2) thinktalk”类型的视频记录了导师在办公室讲课的过程,教授预先准备好的 学习材料(3)数字表文件格式,流行于可汗学院与动画数字表教学(4)传统的幻灯片演示Guo, Kim, and Rubin的研究提供了以下重要发现:(1)较短的视频对学生更有吸引力。视频的最
13、正确长度已经确定为6分钟(3分钟 和9分钟仍然可以接受)。视频越长,花在上面的时间就越少。(2)令人惊讶的是,学习者发现这些视频更有趣,不仅能看到演示文稿,还能 看到导师。有目的的记录导师和演示,打破单调。(3)目标观众更喜欢自制的视频,而不是专业的录音室录音(4) 一个类似可汗学院数字表格的环境比简单的演示更受欢迎。(5)并非为MOOC课程而录制的课程录音根本不受欢迎。(6)此外,更多的支持录音,其中讲师是熟练的演说家,可以吸引观众的注意 力。(7)像视频这样的讲座和演示文稿之间也有质的区别,也就是所谓的教程视频, 因为讲座通常只看一次,而演示文稿那么要看好几次。人们已经确定,视觉信息和音频
14、信息是相互增强的,但视频仍然具有按时间 顺序显示操作的质量,模拟用户在自己的设备上看到这些操作的方式。这种方法 创立了一个所谓的易于遵循的模型。Van der Meij和Van der Meij对五年级和六 年级的学生(平均年龄:1L8岁)进行了一项研究,共有111名儿童(56名男孩和56 名女孩)。建立四组,其中两组采用混合方法(视频和纸质),另两组分别使用视频 或纸质学习活动,然后进行测试。通过纸质材料学习的学生考试成绩为65.4%, 而通过视频学习的学生平均成绩为89.7%o后一组似乎也获得了更深更持久的知 识(76.8%至 55.6%)o总的来说,传播媒体的出现并没有改变人们使用学习平
15、台的方式,因为它们 与已经存在的学习平台很相似,例如,在电视广播的初期,数字材料只不过是录 制的广播节目,而数字书籍那么是扫描书籍。这种传统的教学方法在录制第一批教 学视频时得到了保存。他们后来改进了添加更多的创造力,以得到最大限度地利 用这个广泛传播的媒体。唐麦金托什表示,视频可以提高学习动力的有效性, 但必须彻底检查的价格/效益比。到目前为止,我们已经处理了所谓的按需视频存储在本地并可从互联网上获得, 但是有必要提到视频会议作为一种实时视听内容,这也适用于教育环境。虽然网 络摄像头只能转发视觉信息,但视频会议上的设备配备了更广泛的互动可能性, 可以更成功地将参与者连接起来。阿萨班卡大学在2
16、002年进行的研究揭示了该系统的一些弱点,即软件支持 和带宽;此外,他们在测试免费服务和产品时发现,由于互联网连接速度较慢,在 转发视频和音频内容时出现了大量功能错误。我们不应该忽视这样一个事实:一 些可付费软件也出现了类似的错误。如今,大多数用户都能接入宽带互联网,软 件也在教育领域得到了广泛的开展。in.视频记录仪模块的技术背景我们的课题组和Szeged大学软件工程系的团队准备了一个基于Moodie的 系统,该系统能够创立一个学习者活动的日志文件,并将其保存下来,供以后进 一步研究和评估。系统使用JavaScript解决方案,在用户id、发起的操作类型、 活动确实切web位置和时间的帮助下
17、记录单个用户活动(单击、下载、鼠标移动、 视频视图等)。Esztelecki Peter(2016)的一项研究可能会提供一些与该主题相关的 进一步信息。在同一研究中发表了一篇关于教育大数据的论文。我们使用Moodie作为一个扩展了自己模块的电子学习框架。其中一个模块 负责日志记录,它由两局部组成:客户端局部和服务器端局部。它们都是用 JavaScript实现的:在第一个例子中,我们使用jQuery,在第二个NodeJS中使用 Express。ClientClientServer图1记录视频事件代码的客户端局部处理所有JavaScript事件,处理并将它们存储在一个缓冲 区(事件缓冲区)中。如果
18、缓冲区已经满了,或者用户关闭了页面,或者它处于被 动状态5分钟,那么它将通过 请求将缓冲区的内容发送到服务器端局部。 服务器端局部非常简单:它只是将给定的事件存储到Mongo数据库中,以便进一 步处理。我们的系统支持两种视频:html5视频元素和嵌入式Youtube视频: 系统能够通过捕获JavaScript事件,并以JSON格式处理和存储它们来记录这两 个事件。例如,这里有一个视频播放事件:(type: videoPlay,data: actualTime: 95.458972, videold: video totalTime: 956.48,src: f7eleamingdata/okt
19、atok/sd/13.mp4ntime: ”2016.07.18. 09:04:57.884”,page: suserid: 942 事件有一个类型参数,参数如下:(1) userid:执行操作的用户的ID,如果有匿名用户,那么为0time:活动的日期(3) page:事件发生所在页面的URLtype:事件的类型(参见下面)(4) data:可选参数,取决于事件的类型。(6)对于html5视频元素,系统捕获以下JavaScript事件进行处理:a) play:在本例中,将生成一个videoPlay事件,其参数如下所示:actualTime、 totalTime videoSrc 和 video
20、Idoseeked:系统生成一个 videoSeek 事件,参数为:seekTime, totalTime, videoSrc, videoIdob) pause: 生成一个 videoPause 事件。参数:actualTime, totalTime, videoSrc, videoIdowebkitfullscreenchange mozfullscreenchange fullscreenchange: 它表示全 屏或全屏关闭事件。参数:actualTime, totalTime, videoIdoc) volumechange:它生成一个volumeChange事件,该事件具有与上面相
21、同 的参数,并且具有一个newVolume参数。在使用Youtube的JavaScript API嵌入Youtube视频的情况下,系统捕获以下事 件并生成参数为:actualTime、videold、totalTime和src的事件:(1) onStateChange:取决于给定的事件对象。(2) YT.PlayerState.PLAYING: youtubePlayYT.PlayerState.ENDED: youtubeEnd(3) YT.PlayerState.PAUSED: youtubePauseYT.PlayerState.BUFFERING: youtubeBuffering(4
22、) onPlaybackQualityChange: youtubeQualityonPlaybackRateChange: youtubeRate例如,我们提供了 YT.PlayerState.PLAYING的结构代码:function onPlayerStateChange(event) if (event.data = YT.PlayerState.PLAYING) pushEvent( type: !youtubePlay data: actualTime: event!targetf.getCurrentTime()9 videold: event,target, a id,tota
23、lTime: eventfarget1.getDuration()5 src: event,target, .getVideoUrl()这些事件存储在“事件缓冲区”中(使用pushEvent方法),然后提交给服务器, 以便写入数据库。IV .试验课程和记录数据我们的团队为西格德大学的学生开设了试点课程,涉及两个主题:有意识和 安全的互联网使用和数据库。为了建立一个稳定可靠的测井系统,我们尽了最大 的努力来修复试验研究中可能出现的任何错误。即使在测试阶段,我们也能够积 累一个1.2 GB的日志文件,其中记录了 450多万行原始数据。必须指出的是, 除了从这两项试点研究中收集数据外,其他日志也被保
24、存了下来,这些日志来自 该大学宣布并执行的Moodie课程。在JSON文件的评估过程中,我们团队发现了一个系统错误,与填写测试和 播放视频的过程有关,即没有记录这些事件。后来,这个错误被修复,允许我们 在将来的实时课程会话中将这些数据保存到日志文件中。很快我们就发现,尽管没有使用数据挖掘算法对原始数据进行分析,但是必 须对原始数据进行适当的清理和更改。出于这个原因,EszteleckiPeter开发了一 个基于PHP的程序,该程序将json文件转换为csv(逗号分隔值)文件,csv是一 种数据库管理和数据挖掘工具都可以轻松处理的格式。该程序允许用户选择将哪 些课程或课程数据转换为csv文件;此
25、外,还可以决定将哪些事件放入输出文件。 转换过程生成一个stats文件,该文件指示从事件处理了多少数据,以及转换数 据将花费多少时间,等等。要转换一个L2 GB的文件,其中包含所有课程的信 息和他们在Intel (R) Core (TM) i7-2630QM上的活动,使用2.00GHz处理器和 6.00 GB RAM,需要6分钟。如果SSD系统替换当前的5400 RPM SATAII硬盘 组成,那么这个时间量可以进一步减少,因为大局部消耗的时间都花在硬盘进程 上(编写和重新记录数据)。采用数据挖掘的方法对原始数据进行处理,获取新的信息片段,从而更准确 地映射学习者的行为和成就。我们将进一步将参
26、与者分类,并将他们分组,创立 一个小组,让学生在更个性化的环境中学习,然后建立一个适应学习的平台。使用回归计算,可以在学习者例程的基础上,高精度地预测参与者的结果结 果。同样的过程也可以使用统计匹配来完成,该匹配将一个或多个学习者从数据 库分配到适当的主题。我们未来的计划是将我们的研究课程转移到匈牙利南部的 中学和塞尔维亚的伏伊伏丁那,以便将参与者扩大到几千人,使分析更大量的数 据成为可能。V .结论广播媒体为公众教育提供广泛的学习机会;然而,必须指出的是,视听内容 的出现意味着远程学习领域的一个巨大飞跃。另一个伟大的成就是互联网的普及, 它带来了 Web 2.0提供的工具。最后,视频提供网站
27、的建立和推广也必须提到。 Guo, Kim and Rubin用不同的方法记录他们的教育视频,后来得出结论,较短的 视频对学习者更有吸引力;此外,他们更喜欢导师和演示相互跟随的讲座。然而, 这在很大程度上取决于教师的沟通技巧和授课方式。作者还意识到,录音室录音 不如自制录音受欢迎;此外,学习者不珍惜课堂录音,而这些录音显然不是为 MOOC平台准备的。我们团队创立的日志系统能够在虚拟学习环境中全面重构 学习者的活动,为数据挖掘工具分析学习者的行为和学习风格开辟了领域。我们认同Mark Van Rijmenam的期望,即大数据分析将彻底改变学生学习和 教师教授的方式。of the most con
28、venient time to acquire knowledge. In order to develop E-learning, not only the promotion of LLL (lifelong learning) was crucial, but it was necessary to promote quality education, to advocate international education; though the most important factor was the accelerated progress of information techn
29、ologies.Another notion has also gained ground that advocates learning through mobile devices, the so called Mleaming. It emphasizes that knowledge is not only dependent on time but can be acquired regardless of location. Such technological advancements could simplify the achievement of lifelong lear
30、ning, since advanced studies will become accessible side by side with full- or part-time jobs. An increasing number of universities launch online courses, which have penetrated even to secondary and primary schools where teachers incorporate experimental online courses in their curriculum.We may dif
31、ferentiate between virtual and mixed methods. Sharing learning materials and communication with this virtual method comes off in the virtual learning environment (VLE) with the help of sharing messages and participating in forum discussions. Thus, learners do not meet with their tutors or with each
32、other, which is not a problem in adult education, but is not convenient fbr the K-12 age-group that cannot neglect supervision and meeting peers fbr mutual learning. In this latter case a mixed method, Blended learning, would be more beneficial, since it would imply classical school lessons within a
33、n accepted curriculum, however some elements still can be allowed to be acquired through online methods and individual time-tables regardless of location. Blended learning makes it possible fbr learners to deepen their knowledge that can be measured by testing and seems to be more durable.IL AUDIO-V
34、ISUAL CONTENTS IN ONLINE COURSESAudio-visual contents target listening and visual capacities in order to trigger information acquisition. Such tools and learning materials appeared after the I. World War as object lessons, though in those times slideprojectors were already in use accompanied by a re
35、corded audio tape. The mass popularization of the television in the 60s and 70s peaked just in those times on the Hype cycle representing increased expectations. Most of the studies conducted in the United States tried to prove that education through watching television is more efficient than tradit
36、ional methods, but it was quickly revealed that a single most effective system does not exist. Later, as personal computers appeared in households and Internet coverage spread out covering the entire globe, new opportunities came to life in the 90s, when viewing and sharing video materials helped es
37、tablishing the first online courses.First generation digital learning materials were nothing else than scanned traditional school books and digitalized educational videos, while second generation tools were put together exclusively using digital media and content optimized fbr computer use. Criteria
38、 for the third generation digital educational materials were the following: interactivity, inclusion of a wide range of motion pictures and audio recordings, the use of links for enhanced navigation, methodological guidance, all of which created an independent learning material. The fourth generatio
39、n puts emphasis on interactivity and collaboration whose products did not originate from a single author, but could be derived from various multicultural communities. Real breakthrough in this field was noticeable after the launch of online video sharing sites, like YouTube, since thereafter educati
40、onal videos were easily accessible. For instance, video shares by Khan Academy were visited 663 million times on YouTube since 2006. The expression MOOC (Massive Open Online Course) became wide known around this period which comprises web-based courses with unlimited participation and accessibility
41、fbr free. We can differentiate between four types of educational videos: (1) classroom lessons, which are basically records of school lessons or university lecturesthink talk type recorded videos during which tutors give lectures from their office teaching a pre-established learning material(2) digi
42、tal table file format, popular at Khan Academy with animated digital tables fbr teachingtraditional PowerPoint presentationsThe studies of Guo, Kim, and Rubin provide an insight into the following important findings:(1) Shorter videos are more appealing to students. The optimal length of videos have
43、 been established at 6 minutes (3 and 9 minute are still acceptable). Longer videos lower the amount of time spent on them proportionally with their length.(2) Learners, surprisingly, found those videos more interesting that visualize not only presentations but also tutors. It is purposeful to recor
44、d the tutor as well as the presentation by turns to break monotony.(3) The target audience prefers homemade videos that seem more personal than professional studio recordingsAn environment that resembles to digital tables of the Khan Academy is more popular than simplistic presentations.(4) Recordin
45、gs from school lessons that were not intended fbr a MOOC course are not popular at all.(5) Also more support is given to recordings in which lecturers are skilled orators and can catch the attention of the audience.(6) There is also a qualitative difference between lecture like videos and presentati
46、ons, so called tutorial videos, since lectures are usually watched only once, while presentations are consulted several times.It has been established that visual and audio information strengthen each other, still videos have a quality of showing operations in a chronological order simulating how a u
47、ser would see them on their own device. This approach creates a so called easy-to- follow model. Van der Meij and Van der Meij carried out a study measuring fifth and sixth grade learners ( average age: 11.8 years), altogether 111 children (56 boys and 56 girls). Four groups were established of whom
48、 two groups worked following a mixed method (videos and paperbased), while the other two groups respectively used videos or paper-based activities exclusively for learning, after which they were given a test. Those students who were learning with the help of paper-based materials achieved a test score of 65.4%, while those who were learning by consulting videos, on average, scored 89.7%. The latter group seems to have acquired deeper and long lasting knowledge also shown b