基于相似消息的流行度预测方法

社交网络中消息的流行度预测问题在很多应用领域都有着重要意义。传统的流行度预测方法包括基于特征的方法和基于点过程的方法。基于点过程的方法无法利用历史消息的信息,而基于特征的方法则使用一个统一的模型来对所有的消息进行预测,没有考虑消息的特异性。因此,该文提出了一种基于相似消息的流行度预测方法。对于待预测微博,我们从历史消息选取出与之最相似的前K条消息来进行预测。在计算消息相似度时,我们借助了文档建模领域的LDA模型来学习消息的表示。在数据集上的实验结果表明,该方法可以有效发现在传播模式上与待预测消息相似的历史消息,并在流行度预测任务上取得了比对比模型更好的预测效果。     

“I don’t want to miss a thing”: Adolescents’ fear of missing out and its relationship to adolescents’ social needs,Facebook use,and Facebook related stress

This survey study among adolescents (N = 402) investigates an integrative model that examines (1) the mediating role of adolescents’ fear of missing out (FoMO) in the relationships of adolescents’ need to belong and need for popularity with adolescents’ Facebook use and (2) the relationships of adolescents’ FoMO with adolescents’ perceived stress related to the use of Facebook. Structural equation modeling results indicated that an increased need to belong and an increased need for popularity were associated with an increased use of Facebook. These relationships were mediated by FoMO. Increased FoMO was associated with increased stress related to Facebook use. These results emphasize the important role that FoMO plays in adolescents’ media use and well-being.     

一种以负载平衡为目标的分布式VOD系统视频文件存储方法

本文在研究一种分布式视频服务器负载平衡算法的基础上,引入了服务器端视频文件分盘存储的思想.通过仿真实验分析表明,对视频文件分盘存储能够降低系统的服务拒绝率,更进一步地提高了视频服务器效能.     

一种基于流行度分类特征的托攻击检测算法

基于协同过滤的推荐系统容易受到托攻击的危害, 如何检测托攻击成为推荐系统可靠性的关键. 针对现有托攻击检测手段使用基于评分的分类特征易受混淆技术干扰的局限, 本文从用户选择评分项目方式入手, 分析由此造成的用户概貌中已评分项目的流行度分布情况的不同, 提出用于区分正常用户与虚假用户基于流行度的分类特征, 进而得到基于流行度的托攻击检测算法. 实验表明该算法在托攻击检测中具有更强的检测性能与抗干扰性.     

面向高能物理分级存储的文件访问热度预测

高能物理计算是典型的数据密集型计算,其主要采用基于文件的分级存储方案,根据访问热度的不同将数据存储于不同性能的存储设备上,然而当前数据热度预测采用基于人工经验的启发式算法,准确率较低。提出一种借助长短期记忆网络预测文件未来访问热度的方法,包括网络结构设计、训练和预测算法等。该方法通过划分动态时间窗口构造文件访问特征的时序序列,预测不同数据的访问趋势。在LHAASO高能物理实验数据集上的实验结果表明,与SVM、MLP等算法相比,该方法预测准确率提升了30%左右,具有更强的适用性。     

推荐系统中谁可以协同新用户？

协同过滤作为被成功应用于推荐系统的技术之一,得到了各领域学者的关注。然而随着网络平台新用户和项目的不断增加,协同推荐面临严重的“冷启动”问题的挑战。首先基于用户流行度和长尾分布建立用户推荐能力的度量方法,然后利用用户推荐能力筛选出一个用于推荐的全局核心用户子集,来解决推荐系统的“冷启动”问题。实验结果显示,将构建的全局核心用户集合用于协同推荐,在不降低推荐效果的基础上,可显著降低寻找相似用户的时间复杂度,因而可以将其用于解决推荐实时性问题。     

Understanding adolescent students’ use of Facebook and their subjective wellbeing: a gender-based comparison

Adolescence is marked by many changes, and adolescents seek peer support to compensate for the concomitant psychological and social challenges they face. In this respect, Facebook has become one of the primary tools adolescents use to build and maintain social relationships. The current study examines the determinants of Facebook use and its relationship with adolescents’ subjective wellbeing, stratified by gender. Data were collected from a nationally representative sample of 1,121 adolescent students (aged 12–17 years) consisting of 625 boys and 586 girls. The results indicated that the total number of Facebook friends, the need to belong and the perceived waste of time were correlated with adolescents’ use of Facebook; in turn, Facebook use was positively related to the subjective wellbeing of both male and female students. Gender differences were also observed; the need to belong and the positive relationship between Facebook use and subjective wellbeing were both stronger for male students than female students.     

FPTC:一种信息中心物联网缓存策略

信息中心网络(information-centricnetworking,ICN)作为一种新型未来互联网体系结构应运而生,并广泛应用于物联网领域,其中,缓存技术作为ICN的显著特征,对信息中心物联网的内容传输性能具有重要影响.由于信息中心物联网具有数据频繁更新、用户对数据新鲜度有严格要求等特性,致使传统信息中心网络缓存技术面临挑战.提出一种基于内容流行度和网络拓扑的分布式缓存策略,同时考虑内容新鲜度,各缓存节点通过优先缓存流行度较高且较靠近用户的内容,以最大化缓存效率.为适应物联网内容的频繁更新,提出一种基于灰色预测的内容缓存收益预测方法,便于快速获取新内容的缓存收益值.同时,该策略具有较低的时间和空间开销.仿真实验结果表明:所提方案相比于传统缓存策略,能有效提高缓存效率和命中率,并降低访问延迟,改善用户体验.     

分布式VoD系统中一种分层的缓存管理策略

相对于传统网络服务系统的客户端/服务器构架,分布式VoD系统通过系统中节点问的协作能够获得更好的性能.而节点的存储空间和网络带宽是有限的,因此缓存管理的研究对流媒体服务系统的研发具有重要的意义.该文基于一种灵活的热度估计方法对系统中的存储资源进行分层分析,给出一种减小服务器磁盘I/O负载的优化式及其次优解,同时针对客户节点的不稳定特性提出一种以二次规划为基础的数据置换算法.仿真实验结果表明热度的计算能够较好的反映用户行为的变化,而缓存管理策略有利于减小服务器磁盘数据的读写以及增加系统的可利用资源.     

Semi-supervised community detection on attributed networks using non-negative matrix tri-factorization with node popularity

The World Wide Web generates more and more data with links and node contents, which are always modeled as attributed networks. The identification of network communities plays an important role for people to understand and utilize the semantic functions of the data. A few methods based on non-negative matrix factorization (NMF) have been proposed to detect community structure with semantic information in attributed networks. However, previous methods have not modeled some key factors (which affect the link generating process together), including prior information, the heterogeneity of node degree, as well as the interactions among communities. The three factors have been demonstrated to primarily affect the results. In this paper, we propose a semi-supervised community detection method on attributed networks by simultaneously considering these three factors. First, a semi-supervised non-negative matrix tri-factorization model with node popularity (i.e., PSSNMTF) is designed to detect communities on the topology of the network. And then node contents are integrated into the PSSNMTF model to find the semantic communities more accurately, namely PSSNMTFC. Parameters of the PSSNMTFC model is estimated by using the gradient descent method. Experiments on some real and artificial networks illustrate that our new method is superior over some related stateof- the-art methods in terms of accuracy.