A systematic approach to normalization in probabilistic models

Every information retrieval (IR) model embeds in its scoring function a form of term frequency (TF) quantification. The contribution of the term frequency is determined by the properties of the function of the chosen TF quantification, and by its TF normalization. The first defines how independent the occurrences of multiple terms are, while the second acts on mitigating the a priori probability of having a high term frequency in a document (estimation usually based on the document length). New test collections, coming from different domains (e.g. medical, legal), give evidence that not only document length, but in addition, verboseness of documents should be explicitly considered. Therefore we propose and investigate a systematic combination of document verboseness and length. To theoretically justify the combination, we show the duality between document verboseness and length. In addition, we investigate the duality between verboseness and other components of IR models. We test these new TF normalizations on four suitable test collections. We do this on a well defined spectrum of TF quantifications. Finally, based on the theoretical and experimental observations, we show how the two components of this new normalization, document verboseness and length, interact with each other. Our experiments demonstrate that the new models never underperform existing models, while sometimes introducing statistically significantly better results, at no additional computational cost.     

Instructors' and Students' Perspectives of Student Nagging: Frequency,Appropriateness, and Effectiveness

Nagging is a persistent tactic students use to persuade instructors. The purpose of this study was to compare student (n = 280) and instructor (n = 97) perceptions of student nagging in the classroom, and to examine the relationship between student nagging and instructor satisfaction. Results show that instructors perceive students as using four nagging strategies with significantly greater frequency than students do. In addition, students perceive all seven nagging strategies to be significantly more effective and appropriate than instructors. Finally, there is a significant relationship between particular nagging strategies and instructor satisfaction.     

基于N-Gram的专业领域中文新词识别研究

以植物学作为专业领域的样本,对专业领域的新词自动化识别进行探索。研究选取《中国植物志》作为样本集,在ICTCLAS切词的基础上采用N-Gram统计的方法提取新词的候选项,然后分别按照词频(TF)、文档频率(D)和平均词频(TF/D)对新词候选项排序,取一定范围内的候选项作为识别出的新词。实验结果表明,词频TF筛选新词候选项的识别效果最好,F值为0.65。该方法能够自动产生专业领域的用户词典,具有较强的可移植性。     

Query reformulation mining: models,patterns, and applications

Understanding query reformulation patterns is a key task towards next generation web search engines. If we can do that, then we can build systems able to understand and possibly predict user intent, providing the needed assistance at the right time, and thus helping users locate information more effectively and improving their web-search experience. As a step in this direction, we build a very accurate model for classifying user query reformulations into broad classes (generalization, specialization, error correction or parallel move), achieving 92% accuracy. We then apply the model to automatically label two very large query logs sampled from different geographic areas, and containing a total of approximately 17 million query reformulations. We study the resulting reformulation patterns, matching some results from previous studies performed on smaller manually annotated datasets, and discovering new interesting reformulation patterns, including connections between reformulation types and topical categories. We annotate two large query-flow graphs with reformulation type information, and run several graph-characterization experiments on these graphs, extracting new insights about the relationships between the different query reformulation types. Finally we study query recommendations based on short random walks on the query-flow graphs. Our experiments show that these methods can match in precision, and often improve, recommendations based on query-click graphs, without the need of users’ clicks. Our experiments also show that it is important to consider transition-type labels on edges for having recommendations of good quality.     

Does Online News Reading and Sharing Shape Perceptions of the Internet as a Place for Public Deliberations?

With the rapid development of interactive communication technology, the Internet is a major source of news and also plays an important role in connecting individual members of society. However, Internet users may have different perspectives on whether the Internet positively functions as a medium for public deliberation. Based on the assumption that being exposed to information on public affairs is a crucial step for one's civic engagement, this study explores how individuals’ motives related to news consumption, elaborative online news reading, and online news sharing influence their perceptions of online deliberation using structural equation modeling method (N = 998). The study finds significant relationships between news consumption motives and elaborative news reading and sharing behaviors, but only elaborative reading behavior had a significant effect on one's perceptions of online deliberation. The implications of these findings are discussed.     

Motivations Behind Donors' Contributions to Crowdfunded Journalism

Crowdfunded journalism is a new model for funding journalism in which reporters solicit micropayments from readers to finance their reporting. In the present study, we seek to identify the major motivations behind readers' donations to a pioneering crowdfunded journalism website, Spot.Us. Under the theoretical framework of collective action, we explore a total of 9 motivations: altruism, fun, belief in freedom of content, social, community, self-esteem, understanding, image, and family and friends. Combining self-reported survey data (N = 344) with behavioral data on readers' complete donation records obtained from the Spot.Us server, we found that belief in freedom of content, altruism, and contributing to one's community were self-reported to be highly valued by donors, but fun and family and friends motivations were the only positive predictors for actual donation levels. We discuss the implications of these findings on the sustainability of crowdfunding as a business model for news.     

Victor C. Strasburger and Barbara Jan Wilson. Children,Adolescents, & the Media. Thousand Oaks,CA: Sage, 2002, 560 pp., ISBN 9780761921257 (paperback)

Recent research shows parents manifest parental third-person perceptions on behalf of their children; that is, they believe their children are less affected by media sex and violence than other children. This study (N = 171) found parental third-person perceptions for materialism effects of television and parental first-person perceptions for advanced educational effects of public television. Perceptions of materialism effects on one's own and other children predicted parental mediation, whereas perceptions of education effects predicted support for regulations requiring more educational television.     

eHealth literacy and the motivators for HPV prevention among young adults in Kenya

ABSTRACT

The human papillomavirus (HPV) is one of the most common but preventable sexually transmitted infections that affect males and females but knowledge about it and vaccine uptake remain low. As organizations adopt eHealth communication strategies to communicate about HPV, this study examines the association between eHealth literacy (and related elements) and the motivators for HPV prevention that are based on the protection motivation theory. A survey was conducted among 472 young adults in Kenya and results show moderate eHealth literacy (M = 3.21, SD = 1.03) and online information seeking (M = 3.57, SD = 1.08) but low HPV knowledge and risk perception. eHealth literacy was correlated with HPV knowledge, perceived risk, self-efficacy, and response efficacy whereas online health information seeking were correlated with perceived seriousness of HPV infection. This study suggests examining the discrepancy between online information seeking and knowledge and the disparities in eHealth literacy among young adults in HPV communication.     

基于深度神经网络的微博图书名识别研究

[目的/意义] 微博作为一种新兴的社交媒体平台,被互联网用户广泛关注。微博数据中包含着大量的用户信息、用户行为及用户生成内容,基于微博内容自动识别图书名有利于分析用户阅读兴趣、收集用户对图书的评价和挖掘图书相关知识。[方法/过程] 基于微博的数据特点,提出一种基于深度神经网络的表示学习方法,利用微博中候选图书名的上下文连续向量化表示,实现微博内容中的图书名自动识别。[结果/结论] 实验结果表明,所提出的方法显著优于传统基于特征工程的有指导机器学习方法,并达到91.92%的精确率。     

Using rating matrix compression techniques to speed up collaborative recommendations

Collaborative filtering is a popular recommendation technique. Although researchers have focused on the accuracy of the recommendations, real applications also need efficient algorithms. An index structure can be used to store the rating matrix and compute recommendations very fast. In this paper we study how compression techniques can reduce the size of this index structure and, at the same time, speed up recommendations. We show how coding techniques commonly used in Information Retrieval can be effectively applied to collaborative filtering, reducing the matrix size up to 75 %, and almost doubling the recommendation speed. Additionally, we propose a novel identifier reassignment technique, that achieves high compression rates, reducing by 40 % the size of an already compressed matrix. It is a very simple approach based on assigning the smallest identifiers to the items and users with the highest number of ratings, and it can be efficiently computed using a two pass indexing. The usage of the proposed compression techniques can significantly reduce the storage and time costs of recommender systems, which are two important factors in many real applications.     

A nonparametric term weighting method for information retrieval based on measuring the divergence from independence

In this article, we introduce an out-of-the-box automatic term weighting method for information retrieval. The method is based on measuring the degree of divergence from independence of terms from documents in terms of their frequency of occurrence. Divergence from independence has a well-establish underling statistical theory. It provides a plain, mathematically tractable, and nonparametric way of term weighting, and even more it requires no term frequency normalization. Besides its sound theoretical background, the results of the experiments performed on TREC test collections show that its performance is comparable to that of the state-of-the-art term weighting methods in general. It is a simple but powerful baseline alternative to the state-of-the-art methods with its theoretical and practical aspects.     

Media Substitution in Cable Cord-Cutting: The Adoption of Web-Streaming Television

This study sought to better understand what factors best predict consumers’ intention to cut the cord on cable television and adopt video streaming as their primary source of television. Utilizing media substitution theory as the conceptual framework, this study conducted a nationwide survey (N = 200). Findings show that perceived advantages of streaming applications over traditional television best predicted intentions to cut the cord on cable and adopt Web streaming; these perceptions mediated the relation between user frustrations with using older television technology and intentions to cut the cord. Entertainment needs were not significant predictors of cord-cutting intentions.     

Improving document clustering using Okapi BM25 feature weighting

We investigate the effect of feature weighting on document clustering, including a novel investigation of Okapi BM25 feature weighting. Using eight document datasets and 17 well-established clustering algorithms we show that the benefit of tf-idf weighting over tf weighting is heavily dependent on both the dataset being clustered and the algorithm used. In addition, binary weighting is shown to be consistently inferior to both tf-idf weighting and tf weighting. We investigate clustering using both BM25 term saturation in isolation and BM25 term saturation with idf, confirming that both are superior to their non-BM25 counterparts under several common clustering quality measures. Finally, we investigate estimation of the k1 BM25 parameter when clustering. Our results indicate that typical values of k1 from other IR tasks are not appropriate for clustering; k1 needs to be higher.     

社会化问答平台用户体验影响因素实证分析——以知乎为例

[目的/意义]研究社会化问答平台用户体验的影响因素,为社会化问答平台改善用户体验,提高用户留存率、活跃度提供参考。[方法/过程]构建一个综合模型,以"知乎"用户为调查对象,采用回归分析法对其用户体验的影响因素进行量化分析。[结果/结论]视觉吸引力和需求满足会显著影响社会化问答平台的用户体验;主观情感作为中介变量,对视觉吸引力、内容需求、交互体验与用户体验之间的关系起调节作用;除使用频率与用户年龄外,其他环境变量对用户体验的影响可以忽略不计。     

A living lab as socio-technical ecosystem: Evaluating the Korean living lab of internet of things

Living Lab approaches, as one of socio-technical approaches, are effective strategy for user-driven technology development. The recent development of the Internet of Things (loT) and its various technologies appear to be promising possibilities to adopt Living Lab innovation into community domains. Using Living Lab frame, this study examines the developmental processes of the IoT from a multi-level analysis: a micro approach of user acceptance and experience of IoT services; a meso approach of socio-technical evaluation of selected site; and a macro approach of regulation and strategies on IoT. Through the multi-level approach, it conceptualizes a Living Lab process to create, test, and adept a social IoT environment. The findings should guide governments' promotion of IoT services to increase user acceptance by enhancing usability and benefits and ensuring sustainability. The findings also provide guidelines, strategies, and best practices for practitioners to integrate IoT into communities and society effectively. The insights help to conceptualize how the IoT can be situated and contextualized within human-centered contexts. The results of this study show that creating IoT innovations require prudent coordination of different stakeholders and roles across innovation cycle. Particularly user-centered approach warrants a new innovative way to structure and facilitate user involvement within the context of Living Labs for IoT.     

Advocating for controversial issues: The effect of activism on compliance‐gaining strategy likelihood of use

Previous research has demonstrated that compliance‐gaining strategies can he arrayed on an unidimensional continuum, anchored by positive, pro‐social strategies and negative, antisocial strategies (Hunter & Boster, 1987; Rolojf & Barnicott, 1978, 1979). Individual differences, including argumentativeness and verbal aggression (Boster & Levine, 1988), have been shown to affect one's ethical threshold. This ethical threshold is consistent with a Guttman simplex and determines both how many, as well as how antisocial, compliance‐gaining strategies individuals may consider using. Activists, who are highly involved and committed to their cause and who tend to be less verbally aggressive and more argumentative than the general public (Campo, 1999), are likely to have a lower threshold and to use more strategies. This study examines to what extent activism predicts one's ethical threshold, when controlling for the known effects of argumentativeness, verbal aggressiveness, and sex. Participants (N = 454) at two separate universities completed a repeated measures questionnaire that asked them to indicate their likelihood of use of 25 separate compliance‐gaining strategies. Based on respondents’ likelihood of use of these 25 strategies, the data were consistent with a Guttman simplex. A multiple regression model was tested, and indicated that one's ethical threshold is directly predicted by level of activism and verbal aggressiveness. Sex and argumentativeness were not found to be significant predictors in this revised model. In addition, activists used more strategies than non‐activists. Implications for advancing research related to activism and compliance‐gaining strategy selection are discussed.     

Assessing the Role of Information-Processing Strategies in Learning From Local News Media About Sources of Social Capital

This study examines the mediating effects of information-processing strategies on the relationship between use of local news media, informational use of the Internet, and sources of social capital: interpersonal trust, reciprocity, and associational membership. Analyses of a telephone survey data (n = 546) of American adults show that even though local news media were influential, information-processing strategies were more powerful than attention in explaining learning from local news media about social norms at the individual level. These findings support the usefulness of the cognitive mediation model of information-processing behaviors in examining learning from local news media about social norms. Of the two strategies, elaborative processing played a more important role than active reflection in the mediating process. Informational use of the Internet had a significant and independent effect on associational membership, after demographic, structural anchoring, local media use, and information-processing measures were statistically controlled.     

个性化推荐系统中用户多态聚类研究

针对传统协同过滤算法依赖单一用户需求形态影响推荐效果的问题,提出一种基于用户多态聚类的数字图书馆个性化推荐方法。该方法以改进的海明距离计算候选邻居集,结合多态相似度进行二次聚类,预测用户的多态需求度并形成推荐。实验表明,使用多态聚类产生的推荐精确度上优于单一聚类产生的推荐。     

Scale and Translation Invariant Collaborative Filtering Systems

Collaborative filtering systems are prediction algorithms over sparse data sets of user preferences. We modify a wide range of state-of-the-art collaborative filtering systems to make them scale and translation invariant and generally improve their accuracy without increasing their computational cost. Using the EachMovie and the Jester data sets, we show that learning-free constant time scale and translation invariant schemes outperforms other learning-free constant time schemes by at least 3% and perform as well as expensive memory-based schemes (within 4%). Over the Jester data set, we show that a scale and translation invariant Eigentaste algorithm outperforms Eigentaste 2.0 by 20%. These results suggest that scale and translation invariance is a desirable property.     

基于SEM的信息检索与利用国家精品课程网站用户满意度测评模型实证研究

在前期构建的理论模型和研究假设基础上,以信息检索与利用国家精品课程网站用户为调查对象,搜集286份有效问卷。利用结构方程建模法、SPSS和AMOS软件对通过调查问卷获得的数据进行统计分析,实证检验提出的理论模型和相应的一系列研究假设。研究结果表明用户需要、感知价值、感知质量对用户满意具有较大影响。最后,依据实证分析结果提出几点网站的改进策略。