基于微博用户兴趣话题的相似用户挖掘

相似用户挖掘是提高社交网络服务质量的重要途径，在面向大数据的社交网络时代，准确的相似用户挖掘对于用户和互联网企业等都有重要的意义，而根据用户自己的兴趣话题挖掘的相似用户更符合相似用户的要求。提出了一种基于用户兴趣话题进行相似用户挖掘的方法。该方法首先使用TextRank话题提取方法对用户进行兴趣话题提取，再对用户发表内容进行训练，计算出所有词之间的相似度。提出CP（Corresponding Position similarity）、CPW（Corresponding Position Weighted similarity）、AP（All Position similarity）、APW（All Position Weighted similarity）四种用户兴趣话题词相似度计算方法，通过用户和相似用户间关注、粉丝重合率验证相似用户挖掘效果，APW similarity的相似用户的关注/粉丝重合百分比为1.687%，优于提出的其他三种算法，分别提高了26.3%、2.8%、12.4%，并且比传统的文本相似度方法Jaccard相似度、编辑距离算法、余弦相似度分别提高了20.4%、21.2%、45.0%。因此APW方法可以更加有效地挖掘出用户的相似用户。     

Collective institutional work: the case of Airbnb in Amsterdam,London and New York

ABSTRACT

Given that online platforms disrupt established industries and challenge existing institutions, they can only be successful if their innovation becomes both legal and legitimate. This requires ‘institutional work’ that changes perceptions and regulations within society. Rather than only focussing on the online platform as the sole agent engaging in institutional work, our study analyses institutional work as a collective process. We investigate the case of home-sharing platform Airbnb and the process of institutional change its introduction prompted regarding short-term rental in Amsterdam, London and New York. We find, contrary to the popular view of online platforms as disruptive entrepreneurs, that the platform mainly focusses on creating new institutions rather than disrupting existing ones, and that users and non-users undertake most of the institutional work activities. We also show that different types of actors carry out different types of institutional work suggesting that the process of institutional work is highly distributed.     

为了满足使用者需求而设计——上海地铁研究·使用者需求

在上海地铁研究中,我们发现了许多使用者的不满。这些不满来自于设计师对使用者的缺乏了解和缺乏关注。通过对上海地铁真实的使用者、真实的地点和真实的时间的调研,我们分析出25项上海地铁使用者的需求。并提出希望设计师能够从这些使用者的需求出发,去作为使用者带来良好体验的设计。     

基于多层局部信息融合的在线论坛用户心理危机识别

心理健康问题已经成为当今社会关注的焦点,它严重威胁着家庭和睦与社会稳定.有心理危机的用户经常通过特定的社区论坛或者社交媒体来求助或倾述,这为用户心理危机识别开辟了一个新的途径.论坛帖子长短不一,但判断心理危机的核心信息往往体现在局部内容上,基于此特点,本文构建了一个结合分层长短记忆网络和卷积神经网络的多层局部信息融合模型(Multi-layer Partial Information Fusion model,MPIF),利用论坛用户发布的帖子,检测用户的心理危机严重程度.模型的特点在于:1)利用预训练语言模型BERT对用户帖子中的句子进行向量化表示,充分考虑词语在不同语境中的不同含义表达;2)分别从词、短语、以及句子层面挖掘反映用户心理危机状态的信息,采用深度分层LSTM网络和注意力机制相结合的方式来获取待分类帖子中词语层面以及句子层面的局部信息,利用CNN网络中多种大小不同的卷积核来提取帖子中短语层面的局部信息;3)采用注意力机制和最大池化层,使得模型不仅能够有效地利用局部信息给出心理危机程度的判断,同时可以将这些局部信息展示给心理专家,辅助专家更快了解患者.基于CLPsych2019 Shared Task评测任务的实验结果显示,与评测时排名第一的模型相比,MPIF模型的官方评测指标All-F1值(自杀风险程度a,b,c,d 4个类别的F1值取平均)高出3.9%.经消融实验发现,去除LSTM词语层、CNN短语层、LSTM句子层,All-F1分别下降4%、4.3%、2.4%.     

Leveraging clustering approaches to solve the gray-sheep users problem in recommender systems

Recommender systems apply data mining and machine learning techniques for filtering unseen information and can predict whether a user would like a given item. This paper focuses on gray-sheep users problem responsible for the increased error rate in collaborative filtering based recommender systems. This paper makes the following contributions: we show that (1) the presence of gray-sheep users can affect the performance – accuracy and coverage – of the collaborative filtering based algorithms, depending on the data sparsity and distribution; (2) gray-sheep users can be identified using clustering algorithms in offline fashion, where the similarity threshold to isolate these users from the rest of community can be found empirically. We propose various improved centroid selection approaches and distance measures for the K-means clustering algorithm; (3) content-based profile of gray-sheep users can be used for making accurate recommendations. We offer a hybrid recommendation algorithm to make reliable recommendations for gray-sheep users. To the best of our knowledge, this is the first attempt to propose a formal solution for gray-sheep users problem. By extensive experimental results on two different datasets (MovieLens and community of movie fans in the FilmTrust website), we showed that the proposed approach reduces the recommendation error rate for the gray-sheep users while maintaining reasonable computational performance.     

Adaptive training interfaces for less-experienced,elderly users of electronic devices

A great number of complex electronic devices are now part of our everyday lives. While many of us learn to handle these products by trial and error; others, especially older users with little experience in using electronic devices, need support. In order to allow the user maximum flexibility in terms of learning time and location, a training programme is presented which is implemented as part of the software embedded in the product itself. Particular focus is placed on the effect of adaptive training on learning. In this study, the training versions differed in their ability to adjust their complexity to the user's experience (adaptive user interface complexity) and their capability to support the learner by prompting them during the learning process (adaptive training advice). The results show that the adjustment of complexity had a positive effect on users’ experience: elderly users who trained with an adaptive interface were more successful in learning to use a mobile phone. Adaptive training advice, however, was found to have no significant effects on learners’ success and reduced their self-efficacy. This work offers guidelines on how to design integrated training applications for electronic devices that successfully help elderly users with little prior experience.     

基于改进SVM的互联网用户分类

由于传统模型大量约束样本,导致其学习能力下降,因此设计一个基于改进支持向量机(Support Vector Machine,SVM)的互联网用户分类模型.该模型通过构造样本数据,模拟互联网用户的浏览轨迹;根据用户偏好,制定全新的用户分类策略;基于改进支持向量机,实现对互联网用户的分类.性能测试:3次实验下,此次设计的模型分类准确率平均值为98.56%,超出了预设的期望值,具备分类能力.对比测试:与两组传统用户分类模型相比,此次设计的模型,面对不断增加的样本数据,同样能保持高水平的学习能力.     

Probability-based location anonymity algorithm

As one of the most effective location privacy preservation technologies, the k-anonymity model provides safeguards for location privacy of the mobile client against vulnerabilities for abuse by constructing an anonymous area of k users including the protected one. However, most existing k-anonymity models only utilize the users who are sending requests at recent time. If there are not enough requesting users, the generated anonymous area of the k-anonymity model will be larger than expected. In this paper, a Probability-based Location Anonymity (PLA) algorithm is proposed for protecting location privacy of the mobile users in a road network. The PLA model takes advantage of the historical path track of the users who are not sending the request currently, and then computes the probability into the anonymous section so that it can greatly reduce the size of the anonymous area. Experimental results show that the PLA algorithm is superior to the k-anonymity and it increases its anonymous efficiency enormously.     

非时隙结构认知无线网络中非授权用户传输时延分析

在非时隙结构认知无线网络中,以随机侦听多址方案为基础构建了系统的数学模型,利用柯尔莫哥洛夫费勒前向方程得出系统的稳态概率,并采用拉普拉斯斯帝尔阶斯变换推导出非授权用户数据包的平均传输时延.数值分析表明,信道忙的概率和平均退避时间与传输时延呈正比关系,不同的载波侦听多址方案对传输时延影响很大,其中全侦听多址方案的传输时延最小.     

移动搜索终端用户行为调查研究

移动搜索正在深入到移动互联网用户的生活,为更好地了解移动搜索终端用户行为,以提供更好的服务,对移动搜索终端用户的行为进行了调查。共发出调查问卷522份,收到有效问卷508份,通过对有效问卷进行深入研究,分析移动用户使用移动搜索的行为习惯,得出移动搜索呈现本地化、使用需求碎片化、使用习惯已经形成、用户年龄较为集中等结果,为探索移动搜索新业务提供有效的数据支持,并从搜索精度、搜索方式、提高搜索质量等方面对移动搜索的发展提出合理化建议。