众包直播中精彩片段的自动化识别

近年来,基于众包的视频直播平台逐渐兴起,以其丰富的观众-主播交互机制吸引广大用户观看.针对直播平台的分析也随之成为流媒体服务领域的一个研究热点.直播过程中精彩片段的自动提取对于标签生成、视频分类和内容推荐等方面而言至关重要,然而现有的精彩片段检测大多围绕音频、视频数据本身展开,如视频语义分析、音频情感感知等,缺乏对用户交互属性的合理利用.本文以斗鱼直播平台为例,通过分析观众的发弹幕与送礼物行为,提出了基于直播间弹幕数量时间序列和礼物价值时间序列的精彩片段自动化检测方法.首先利用z-score方法检测序列高潮,然后对高潮做样本标注和特征构建,最后采用随机森林对序列高潮分类并识别内容高潮,即精彩片段.结果表明,模型能够以较高的准确率完成精彩片段的自动化识别任务.     

面向国内直播行业的分布式弹幕爬虫研究

近年来,依托视频行业与直播行业的兴盛,弹幕服务迅速发展。然而主流弹幕服务的弹幕环境一直以来缺乏监管,主播与用户违规行为屡禁不止,对直播弹幕的学术研究稀少,亟需开发针对弹幕的采集处理方案。针对国内知名弹幕服务的技术特征,设计一种分布式直播弹幕爬虫系统方案。分析并提出相应房间连接的建立机制与弹幕采集机制:对开放API的服务直接采用轻量级客户端实现;对基于Adobe Flash且不开放API的服务,用基于Chromium浏览器的Electron模拟浏览直播间网页,并改写其PPAPI插件界面实现,旁路Flash网络流量从而实现抓取。在某知名弹幕平台上进行了验证性实验,表明该系统能够调度IP地址资源进行较大规模抓取,且性能较好,能够处理平均134条每秒、峰值超过1 000条每秒的弹幕流量。     

基于聚类的网络直播群体行为建模分析

近年来,随着互联网技术的不断发展,以及手机、平板电脑等移动终端的普及,网络直播逐渐兴起并壮大.国内众多直播平台基本都有送礼机制,允许观众购买平台提供的虚拟礼物来打赏主播.观众的打赏对于主播和平台来说都是主要的收入来源之一,所以理解观众的行为以挖掘观众的用户价值,提升用户的变现能力就显得尤为重要.本文以斗鱼直播平台为例,聚焦于直播平台上的高消费群体,通过构建观众特征,采用聚类方法分析高消费群体的行为.实验结果表明,高消费观众可被分为特征有明显差异的三类群体.对这三类观众的特征,本文进一步进行详细分析,为直播平台面向用户的差异化产品服务提供依据.     

手机产品的差异性对手机在线阅读行为的影响

本文研究的是手机产品的差异性对手机在线阅读行为的影响.通过分析用户的手机在线阅读行为,从“手机观感”,“手机性能”,“网络情况”三大基本影响因素进行探讨分析,通过相关理论研究以及问卷调查的方式得出手机性能、显示舒适度对手机在线阅读行为的影响最大,以及阅读平台的性能、版面设计、附加功能丰富度对其关系都具有一定的调节作用,且调节作用的大小不一.     

在线社会网络用户的信息分享行为预测研究

在线社会网络中信息的传播路径包含着用户对内容、来源等的偏好信息,研究运用信息的传播路径来预测用户信息分享行为的方法。基于传播路径的信息过滤能力研究了信息在网络中的传播过程和信息传播路径的转换方法。运用基于关联规则的分类算法对在线社会网络中的信息分享行为进行预测。以新浪微博为例对微博用户的转发行为进行了预测,结果表明该方法对在线社会网络中的活跃用户的信息分享行为的预测具有较好的效果。     

在线用户打分行为长记忆效应与信任关系研究

对在线打分行为的动态研究能够帮助深入理解社交网络用户集群行为和信任关系的演化机制,当前许多在线系统用户能够通过对物品进行打分传达自己的观点。通过去趋势波动分析法研究了用户打分行为在信任关系建立前后的长记忆效应,并通过随机化打分时间和信任时间建立零模型,最后进行用户打分行为异质性分析。采用Epinions数据集进行实证研究,结果表明用户打分的长记忆效应在信任关系建立前出现下降趋势（8.06%）,并于之后逐步回升（8.43%）,而在两个零模型中赫斯特指数分别稳定在0.5和0.6左右,且用户长记忆效应变动与用户度呈正相关,Pearson相关系数分别为0.9358和0.9278。该工作有助于深入理解用户集群行为和信任关系的动态演化机制。     

基于云存储的3D直播系统设计与实现

随着互联网技术的快速发展,观看网络直播成为人们主要的线上娱乐方式之一。目前市面上存在很多直播平台,在这些平台日趋广泛的同时,也产生了很多改进的空间,如何增强用户的临场感体验是其中至关重要的一个方面。一般在实际观看中,用户所看到的是2D画面,观看体验感不强,且主播画面与用户实时显示画面,时延大概在2s以上,互动感较差。采用3D直播的方式,将主播以三维虚拟人物形象的方式呈现,将增强用户的体验感。而现有的3D直播平台存在一定的局限性,用户依然是被动的观看直播画面,因此对现有直播平台架构进行了研究和改进。基于Unity软件,采用C/S架构,将人物渲染的流程从局域网拓展到广域网中,在客户端将被动观看变为主动观看,再通过引入CDN和5G网络技术降低传输时延,完成3D直播系统平台设计与搭建。通过该平台可拉近用户与主播的心理距离,增加用户观看网络直播的趣味性与临场感体验。     

在线社会网络的测量与分析

Facebook、Twitter、人人网和新浪微博等社交网站逐渐成为互联网上用户数量最多、最受欢迎的网站.近年来,国内外已有大量研究工作深入考察在线社会网络的拓扑结构和用户行为,这对理解人类的社会行为、改进现有的网站系统和设计新的在线社会网络应用具有重要意义.文中从测量角度对在线社会网络的拓扑结构、用户行为和网络演化等方面进行了综述,总结了常见的测量方法和典型的网络拓扑参数,着重介绍了用户行为特征、用户行为对网络拓扑的影响以及网络的演化.可以看出,随着研究的深入,在线社会网络的新特征逐渐被大家认识和理解,包括好友少的用户的交流范围集中在小部分好友,而好友多的用户联系的好友更均匀;用户之间的交互减小了在线社会网络的聚类系数,使网络结构更松散;边的生成受优先连接和临近偏倚的共同影响;小社团倾向于和大社团合并,大社团倾向于分裂为两个规模相当的小社团等.     

社交网络水军用户的动态行为分析及在线检测

利用社交网络用户的静态行为特征识别水军用户,无法检测水军用户的动态行为且难以应用于在线检测的环境。为此,构造社交网络用户的动态行为特征,分析正常用户和水军用户间的差异,以半监督模型为基础,结合动静行为特征构建在线检测模型,通过静态行为特征聚类及动态行为特征过滤筛选,使半监督模型利用最有价值的未标记用户数据进行增量学习,从而检测水军用户。实验结果表明,该模型的F1值高达93.33%,平均训练时间约为2 min,能够有效检测社交网络上的水军用户。     

网络号百用户兴趣模型挖掘算法

为了实现在线推荐信息服务,要对网络号百用户的访问行为进行分析,荻取用户访问聚类模型,从而在聚类模型的基础上进行在线推荐.介绍获取用户访问路径信息的方法,对用户访问路径信息建立相似度矩阵,基于相似度矩阵改进K-means算法,据此进行用户模型聚类,给出分析案例,并说明算法实现过程.     

Incentive Cooperation Strategies for Peer-to-Peer Live Multimedia Streaming Social Networks

Multimedia social networks have become an emerging research area, in which analysis and modeling of the behavior of users who share multimedia are of ample importance in understanding the impact of human dynamics on multimedia systems. In peer-to-peer live-streaming social networks, users cooperate with each other to provide a distributed, highly scalable and robust platform for live streaming applications. However, every user wishes to use as much bandwidth as possible to receive a high-quality video, while full cooperation cannot be guaranteed. This paper proposes a game-theoretic framework to model user behavior and designs incentive-based strategies to stimulate user cooperation in peer-to-peer live streaming. We first analyze the Nash equilibrium and the Pareto optimality of two-person game and then extend to multiuser case. We also take into consideration selfish users' cheating behavior and malicious users' attacking behavior. Both our analytical and simulation results show that the proposed strategies can effectively stimulate user cooperation, achieve cheat free, attack resistance and help to provide reliable services.      

More than watching: An empirical and experimental examination on the impacts of live streaming user-generated video consumption

Live streaming user-generated video (UGV), a nascent format of crowdsourced content, has grown massive popularity among social media users and is believed to have substantial potential business influences. However, industry practitioners express concerns regarding this new form on social media platforms and the influences of live streaming UGV consumption lack research. Motivated such, through a uniquely merged dataset from the video game industry, we conduct a series of panel time-series empirical analyses to investigate the business value of live streaming UGV consumption. Further, we propose a conceptual framework based on the sense of community literature to understand the consequences of live streaming UGV consumption and verify it through several online experiments. This research contributes to the IS literature by shedding light on the influences of emergent crowdsourced content, providing a conceptual framework to existing literature, and offering rich managerial implications and guidelines to managers regarding utilizing live streaming UGVs.     

Energy optimization for mobile video streaming via an aggregate model

Wireless video streaming on smartphones drains a significantly large fraction of battery energy, which is primarily consumed by wireless network interfaces for downloading unused data and repeatedly switching radio interface. In this paper, we propose an energy-efficient download scheduling algorithm for video streaming based on an aggregate model that utilizes user’s video viewing history to predict user behavior when watching a new video, thereby minimizing wasted energy when streaming over wireless network interfaces. The aggregate model is constructed by a personal retention model with users’ personal viewing history and the audience retention on crowd-sourced viewing history, which can accurately predict the user behavior of watching videos by balancing “user interest” and “video attractiveness”. We evaluate different users streaming multiple videos in various wireless environments and the results illustrate that the aggregate model can help reduce energy waste by 20 % on average. In addition, we also discuss implementation details and extensions, such as dynamically updating personal retention, balancing audience and personal retention, categorizing videos for accurate model.     

An efficient hybrid push-pull methodology for peer-to-peer video live streaming system on mobile broadcasting social media

With the rapid growth of wireless communication technology, the availability of highly flexible and video-friendly mobile terminal platforms (such as smartphones and tablets), the emergence of major video content providers (like YouTube, Ustream, and PPTV, which provide a large catalog of attractive contents), Peer-to-Peer (P2P) live video streaming over the wireless and Internet is becoming more and more attractive to users. One of the main challenges is to provide a good quality of service though the dynamic behavior of the network. Traditionally, tree-based model uses a push method, that broadcaster transfers data to other users. This model has low start-up delay. However, there are two main problems in this method: if the bandwidth of an internal node is low, children nodes may lose data and when an internal node failure, other nodes can’t receive data until completing the recovery of the tree. On the other hand, mesh-based model uses a pull method, has low bandwidth of a neighbor node by pulling necessary data from a number of neighbor nodes. However, mesh-based model requires large buffers to support pull data from neighbor peers and there is an adjustment between minimum delay by sending pull request and overhead of whole system. So, both models have their own strengths and weaknesses. This paper proposes a new hybrid push-pull live P2P video streaming protocol called MobileCast that combines the benefits of pull and push mechanisms for live video delivery. We present new topology for P2P network with more stable and provide better video streaming quality. Our main goal is to minimize the network end-to-end delay, startup time, overhead, packet loss compared to the pure mesh networks, pure tree networks and provide a good quality of service though the dynamic behavior of the network.

     

一种基于用户播放行为序列的个性化视频推荐策略

本文针对在线视频服务网站的个性化推荐问题,提出了一种基于用户播放行为序列的个性化推荐策略.该策略通过深度神经网络词向量模型分析用户播放视频行为数据,将视频映射成等维度的特征向量,提取视频的语义特征.聚类用户播放历史视频的特征向量,建模用户兴趣分布矩阵.结合用户兴趣偏好和用户观看历史序列生成推荐列表.在大规模的视频服务系统中进行了离线实验,相比随机算法、基于物品的协同过滤和基于用户的协同过滤传统推荐策略,本方法在用户观看视频的Top-N推荐精确率方面平均分别获得22.3%、30.7%和934%的相对提升,在召回率指标上分别获得52.8%、41%和1065%的相对提升.进一步地与矩阵分解算法SVD++、基于双向LSTM模型和注意力机制的Bi-LSTM+Attention算法和基于用户行为序列的深度兴趣网络DIN比较,Top-N推荐精确率和召回率也得到了明显提升.该推荐策略不仅获得了较高的精确率和召回率,还尝试解决传统推荐面临大规模工业数据集时的数据要求严苛、数据稀疏和数据噪声等问题.     

Enabling broadcast of user-generated live video without servers

We are witnessing the unprecedented popularity of User-Generated-Content (UGC) on the Internet. While YouTube hosts pre-recorded video clips, in near future, we expect to see the emergence of User-Generated Live Video, for which any user can create its own temporary live video channel from a webcam or a hand-held wireless device. Hosting a large number of UG live channels on commercial servers can be very expensive. Server-based solutions also involve various economic, copyright and content control issues between users and the companies hosting their content. In this paper, leveraging on the recent success of P2P video streaming, we study the strategies for end users to directly broadcast their own live channels to a large number of audiences without resorting to any server support. The key challenge is that end users are normally bandwidth constrained and can barely send out one complete video stream to the rest of the world. Existing P2P streaming solutions cannot maintain a high level of user Quality-of-Experience (QoE) with such a highly constrained video source. We propose a novel Layered P2P Streaming (LPS) architecture, to address this challenge. LPS introduces playback delay differentiation and constructs virtual servers out of peers to boost end users’ capability of driving large-scale video streaming. Through detailed packet-level simulations and PlanetLab experiments, we show that LPS enables a source with upload bandwidth slightly higher than the video streaming rate to stream video to tens of thousands of peers with premium quality of experience.     

基于知识图谱和重启随机游走的跨平台用户推荐方法

针对单一社交网络平台中推荐相似用户结果单一,对用户兴趣和行为信息了解不够全面的问题,提出了基于知识图谱和重启随机游走的跨平台用户推荐方法（URCP-KR）。首先,在分割、匹配出的目标平台图谱和辅助平台图谱的相似子图中,利用改进的多层循环神经网络（RNN）预测出候选用户实体,再综合利用拓扑结构特征相似度和用户画像相似度筛选出相似用户;然后,将辅助平台图谱中的相似用户的关系信息补全到目标平台图谱;最后,计算目标平台图谱中的用户游走到社区内每个用户的概率,从而得到用户之间的兴趣相似度来实现用户推荐。实验结果表明,与协同过滤（CF）算法、基于跨平台的在线社交网络用户推荐算法（URCP）和基于多开发者社区的用户推荐算法（UR-MC）相比,URCP-KP在推荐精确率及推荐多样性等方面均有所提高,推荐精确率最高可达95.31%,推荐覆盖率最高可达88.42%。     

一种基于P2P的短视频分享系统

现有在线短视频分享策略通常采用C/S架构,给视频服务器带来较大的带宽压力。为此,提出一种采用点对点方式的在线短视频分享系统IShare,该系统结合用户点播偏好和视频文件之间的社会网络特性实现视频分享。IShare主要包括基于点播兴趣的节点分簇和视频数据源节点的查找2个核心技术。实验结果表明,IShare具备较好的视频数据源节点查找能力,可降低视频服务器带宽资源消耗。     

Structural analysis of value creation in software service platforms

As software service platforms grow in number of users and variety of service offerings, it raises the question of how this phenomenon impacts the value obtained by users. This paper identifies system usability, service variety, and personal connectivity to be the major determinants that contribute to the value offered to users on mobile software service platforms. A structural equation model, which is based on utility theory, technology acceptance theory, and the theory of network externalities, has been constructed from seven observed constructs, reflecting the three determinants and the user value. The lower bound of user value is estimated through the user’s willingness-to-pay for services and the user’s willingness to spend time on using services. For the validation, a co-variance-based structural equation analysis has been conducted on online survey data of 210 users of mobile service platforms (e.g., Android, iOS). The results show that the number of services used and the number of active user connections were found to be the strongest constructs explaining user value. Perceived usefulness did not explain user value as much. In total, they can explain 49 % of the value that the user receives from the platform. The implication of this result is that users’ value from a software service platform cannot be explained by the technology acceptance model itself. Instead, an approach that, as used in this research, of integrating network externality theory, utility theory, and technology acceptance theory is necessary.