查询推荐技术综述

本文介绍了用户搜索中查询推荐技术的相关概念、研究现状;深入分析了目前常见的推荐算法及推荐系统中的隐私保护问题;最后,归纳了查询推荐技术的研究热点。     

协同推荐pLSA模型的动态修正

提出一种对协同推荐的pLSA模型进行修正的算法,在推荐的过程中通过逐步修正PLSA模型中用户的兴趣模型,来适应用户兴趣的变化。通过实验证明该算法能够取得更好的推荐准确度。     

The impact of data obfuscation on the accuracy of collaborative filtering

Collaborative filtering (CF) is a widely-used technique for generating personalized recommendations. CF systems are typically based on a central storage of user profiles, i.e., the ratings given by users to items. Such centralized storage introduces potential privacy breach, since all the user profiles may be accessible by untrusted parties when breaking the access control of the centralized system. Hence, recent studies have focused on enhancing the privacy of CF users by distributing their user profiles across multiple repositories and obfuscating the user profiles to partially hide the actual user ratings. This work combines these two techniques and investigates the unavoidable side effect of data obfuscation: the reduction of the accuracy of the generated CF predictions. The evaluation, which was conducted using three different datasets, shows that considerable parts of the user profiles can be modified without observing a substantial decrease of the CF prediction accuracy. The evaluation also indicates what parts of the user profiles are required for generating accurate CF predictions. In addition, we conducted an exploratory user study that reveals positive attitude of users towards the data obfuscation.     

Analyzing collaborative learning processes automatically: Exploiting the advances of computational linguistics in computer-supported collaborative learning

In this article we describe the emerging area of text classification research focused on the problem of collaborative learning process analysis both from a broad perspective and more specifically in terms of a publicly available tool set called TagHelper tools. Analyzing the variety of pedagogically valuable facets of learners’ interactions is a time consuming and effortful process. Improving automated analyses of such highly valued processes of collaborative learning by adapting and applying recent text classification technologies would make it a less arduous task to obtain insights from corpus data. This endeavor also holds the potential for enabling substantially improved on-line instruction both by providing teachers and facilitators with reports about the groups they are moderating and by triggering context sensitive collaborative learning support on an as-needed basis. In this article, we report on an interdisciplinary research project, which has been investigating the effectiveness of applying text classification technology to a large CSCL corpus that has been analyzed by human coders using a theory-based multi-dimensional coding scheme. We report promising results and include an in-depth discussion of important issues such as reliability, validity, and efficiency that should be considered when deciding on the appropriateness of adopting a new technology such as TagHelper tools. One major technical contribution of this work is a demonstration that an important piece of the work towards making text classification technology effective for this purpose is designing and building linguistic pattern detectors, otherwise known as features, that can be extracted reliably from texts and that have high predictive power for the categories of discourse actions that the CSCL community is interested in.

A domain model of Web recommender systems based on usage mining and collaborative filtering

Considering the increasing demand of multi-agent systems, the practice of software reuse is essential to the development of such systems. Multi-agent domain engineering is a process for the construction of domain-specific agent-based reusable software artifacts, like domain models, representing the requirements of a family of multi-agent systems in a domain, and frameworks, implementing reusable agent-based design solutions to those requirements. This article describes the domain modeling tasks of the MADEM methodology and a case study on the application of GRAMO, a MADEM technique, for the construction of the domain model of ONTOWUM, specifying the common and variable requirements of a family of Web recommender systems based on usage mining and collaborative filtering.     

利用Tri-training算法解决推荐系统冷启动问题

随着社交网络的发展,推荐系统日趋重要,而冷启动问题是推荐系统中的关键问题。设计了一种基于上下文的半监督学习框架TSEL,对矩阵分解模型SVD进行扩充以支持更多形式的上下文信息,利用Tri-training框架训练各个模型。与其他解决推荐系统冷启动问题的半监督方法(如Co- training)相比,该方法有着更好的效果。Tri-training框架能够更加方便地引入更多推荐模型,具有更好的可扩展性。将Tri-training框架加以 扩展,提出了基于用户活跃度生成无标记教学集合的算法和更加丰富的对矩阵分解模型扩充的形式。在真实数据集MovieLens上进行验证,获得了更好的实验效果。     

A design to promote group learning in e-learning: Experiences from the field

The Internet enables learners to be brought together where they can cooperate in learning in groups without space and time limitations. It is, however, quite a challenge to form ideal groups in a short time and ensure satisfactory interaction for learners in cyberspace. In this study, we propose a useful grouping method to help teachers improve group-learning in e-learning by first establishing effective groups with rules based on data mining, and then facilitating student interaction using a system that monitors members’ communication status. Field observations and quantitative evidence show the validity and practicability of the proposed method.     

Creating recommendations on electronic books: A collaborative learning implicit approach

Recommender systems appear among other reasons with the purpose to improve web information overload and ease information recovery. This kind of systems aid users to find contents in a non-difficult way and with minimal effort. Even though, a great number of these systems performance requires contents to be explicitly rated in order to determine user’s interest. When interacting with electronic books this performance may alter users reading and understanding patterns as they are asked to stop reading and rate the content. Therefore, the analysis of user behavior, preferences and reading background can be considered suitable for a recommender system to build collective web knowledge in a collaborative learning context. This way, recommender system can assist users in finding contents of their interest without explicit rating based on previous constructed knowledge. The goal of this research is to propose an architecture to build a content recommendation platform based on eBook reading user behavior, allowing users to learn about the digital content collaboratively. This platform is formed by web readers’ community that aids members in finding contents of their interest in an automatic way and with minimal effort.     

Design of educational multiplayer videogames: A vision from collaborative learning

Videogames and their specific devices can be used to improve learning process since they are very attractive for children. In this way, pupils increase their cognitive skills, the time dedicated to learn, their motivation for learning, their concentration and their attention while they are working/playing. The subject of “learning by play” is behind the introduction of recreational educational techniques in the classroom. If we also consider the increasing presence of new technologies in society in general and in classrooms in particular, we encounter a new way of teaching/learning. Moreover, several studies in the area of computer supported collaborative learning (CSCL) have proved that learning in a group environment (both actively and interactively) is much more productive for pupils than traditional education.Our main objective is to reduce the complexity of introducing the collaborative learning techniques into development of educational videogames. So, in this paper we analyze the use of videogames as a particular case of new technologies in the classroom and we present a set of design guidelines to enable us to incorporate the features of collaborative learning in the videogame development process. We also explore how these guidelines affect the videogame architecture and how they can be applied when designing a videogame. As a practical example of using our proposal we have designed an educational videogame with group activities which aim is to learn the vowels.     

安全推荐系统中基于信任的检测模型

在网格环境中,推荐系统通过提供高品质的个性化推荐,帮助网格用户选择更好的服务。另外,推荐系统也应用于虚拟机管理平台来评估虚拟机的性能和可靠性。然而,推荐结果对用户偏好信息的敏感性使得推荐系统易受到人为攻击(用户概貌注入攻击或托攻击)。本文中,我们提出并评估了一种新的基于信任的安全检测算法以保护推荐系统抵御用户概貌注入攻击。并且,我们分别在用户级和项目级上讨论了信任检测与RDMA检测的结合。最后,我们通过试验表明这些新的安全检测机制可以取得更好的检测精度。     

The multi-user computer-aided design collaborative learning framework

New developments to computer-aided design (CAD) software transform a once solitary modelling task into a collaborative one. The emerging multi-user CAD (MUCAD) systems allow virtual, real-time collaboration, with the potential to expand the learning outcomes and teaching methods of CAD. This paper proposes a MUCAD collaborative learning framework (MUCAD-CLF) to interpret backend analytic data from commercially available MUCAD software. The framework builds on several existing metrics from the literature and introduces newly developed methods to classify CAD actions collected from users’ analytic data. The framework contains two different classification approaches of user actions, categorizing actions by action type (e.g., creating, revising, viewing) and by design space (e.g., constructive, organizing), for comparative analysis. Next, the analytical framework is applied via a collaborative design challenge, corresponding to over 20,000 actions collected from 31 participants. Illustrative analyses utilizing the MUCAD-CLF are presented to demonstrate the resulting insight. Differences in CAD behaviour, indicating differences in learning, are observed between teams made up entirely of novices, entirely of experienced users, or a mix. In pairs of experts and novices, we see both a perceived high-satisfaction apprenticeship experience for the novices and preliminary evidence of an increase in expert design behaviours for the novices. The proposed framework is critical for MUCAD systems to make the most of the educational possibility of combining technical skill-building with team collaboration. Preliminary evidence collected in a fully-virtual design learning activity, and analyzed using the proposed MUCAD-CLF, shows that novice students gain advanced CAD design knowledge when collaborating with experienced teammates. With the user data captured by modern MUCAD software and the MUCAD-CLF presented herein, instructors and researchers can more efficiently assess and visualize students’ performance over the design learning process.     

A collaborative filtering framework based on fuzzy association rules and multiple-level similarity

The rapid development of Internet technologies in recent decades has imposed a heavy information burden on users. This has led to the popularity of recommender systems, which provide advice to users about items they may like to examine. Collaborative Filtering (CF) is the most promising technique in recommender systems, providing personalized recommendations to users based on their previously expressed preferences and those of other similar users. This paper introduces a CF framework based on Fuzzy Association Rules and Multiple-level Similarity (FARAMS). FARAMS extended existing techniques by using fuzzy association rule mining, and takes advantage of product similarities in taxonomies to address data sparseness and nontransitive associations. Experimental results show that FARAMS improves prediction quality, as compared to similar approaches. Cane Wing-ki Leung is a PhD student in the Department of Computing, The Hong Kong Polytechnic University, where she received her BA degree in Computing in 2003. Her research interests include collaborative filtering, data mining and computer-supported collaborative work. Stephen Chi-fai Chan is an Associate Professor and Associate Head of the Department of Computing, The Hong Kong Polytechnic University. Dr. Chan received his PhD from the University of Rochester, USA, worked on computer-aided design at Neo-Visuals, Inc. in Toronto, Canada, and researched in computer-integrated manufacturing at the National Research Council of Canada before joining the Hong Kong Polytechnic University in 1993. He is currently working on the development of collaborative Web-based information systems, with applications in education, electronic commerce, and manufacturing. Fu-lai Chung received his BSc degree from the University of Manitoba, Canada, in 1987, and his MPhil and PhD degrees from the Chinese University of Hong Kong in 1991 and 1995, respectively. He joined the Department of Computing, Hong Kong Polytechnic University in 1994, where he is currently an Associate Professor. He has published widely in the areas of computational intelligence, pattern recognition and recently data mining and multimedia in international journals and conferences and his current research interests include time series data mining, Web data mining, bioinformatics data mining, multimedia content analysis,and new computational intelligence techniques.     

Efficiency of learning environment using GeoGebra when calculus contents are learned in collaborative groups

In this paper we present a modern approach of teaching mathematics based on the computer supported collaborative learning (CSCL) of calculus contents. The collaborative learning was used in calculus course at the University of Novi Sad, Serbia, for examining functions and drawing their graphs. In 2012 the authors decided to improve the collaborative learning introducing GeoGebra application. Small four member groups were formed by using Kagan's (1994) principles. Two groups of students, the experimental, and the control one were observed. The students in the experimental group learned with the help of GeoGebra, and the students in the control group learned without using GeoGebra.Comparison between those two groups of the first year calculus students, regarding their way of learning and the results achieved, is described below. Before the students' collaborative learning, they were tested with a pre-test and their knowledge necessary for examining functions was verified. The pre-test showed that there was no significant statistical difference between the experimental and the control group. The experimental group worked with the help of the computer and the control one without it. After the collaborative learning, the students were tested with a test (colloquium) and the results of the experimental group were significantly better than the results of students in the control group. At the end of the course the students did their exams (post-test), and the results of the experimental group were significantly better than the results of students in the control group.Some students from the experimental group had to answer questions in an interview related to the use of GeoGebra during their collaborative learning. In order to see the students' difficulties in solving problems, students in the experimental group were asked to cross out incorrect parts of solutions, not to erase them. The teachers reviewed the students' tasks done during the collaborative learning and after that the students who had corrected their mistakes were invited for an interview about using GeoGebra for overcoming their difficulties. Based on the students' results in the tests, answers in the questionnaire and in the interview, it can be concluded that GeoGebra has enabled an easier learning of this material. The GeoGebra package enables the students to check whether each step in the process of solving a task was correctly done or not. The results of our research show that GeoGebra can help those students having insufficient knowledge (necessary for solving those tasks) to improve it.We can say that our research shows that the students' learning achievement in examining functions and drawing their graphs is better when they use GeoGebra, working in collaborative groups than without using it. Also, GeoGebra enables creation of effective learning environment for examining functions and drawing their graphs.     

Interactive web environment for collaborative and extensible diagram based learning

Nowadays there is a growing need of ubiquity for learning, research and development tools, due to the portability and availability problems concerning traditional desktop applications. In this paper, we suggest an approach to avoid any further download or installation. The main goal is to offer a collaborative and extensible web environment which will cover a series of domains highly demanded by different kinds of working groups, in which it is crucial to have tools which facilitate the exchange of information and the collaboration among their members. The result of those interactions would be the development of one or several diagrams accessible from any geographical location, independently of the device employed. The environment can be adapted through personalized components, depending on the type of diagram that the user wants to interact with and the users can also create new elements or search and share components with other users of the community. By means of this environment, it will be possible to do research on the usability of collaborative tools for design diagrams, as well as research on the psychology of group interactions, assessing the results coming from the employment of known methodologies, techniques, paradigms or patterns, both at an individual and at a collaborative group level.     

Competitive collaborative learning

Intuitively, it is clear that trust or shared taste enables a community of users to make better decisions over time, by learning cooperatively and avoiding one another's mistakes. However, it is also clear that the presence of malicious, dishonest users in the community threatens the usefulness of such collaborative learning processes. We investigate this issue by developing algorithms for a multi-user online learning problem in which each user makes a sequence of decisions about selecting products or resources. Our model, which generalizes the adversarial multi-armed bandit problem, is characterized by two key features:

(1)

The quality of the products or resources may vary over time.

(2)

Some of the users in the system may be dishonest, Byzantine agents.

Decision problems with these features underlie applications such as reputation and recommendation systems in e-commerce, and resource location systems in peer-to-peer networks. Assuming the number of honest users is at least a constant fraction of the number of resources, and that the honest users can be partitioned into groups such that individuals in a group make identical assessments of resources, we present an algorithm whose expected regret per user is linear in the number of groups and only logarithmic in the number of resources. This bound compares favorably with the naïve approach in which each user ignores feedback from peers and chooses resources using a multi-armed bandit algorithm; in this case the expected regret per user would be polynomial in the number of resources.     

Feature Weighting and Instance Selection for Collaborative Filtering: An Information-Theoretic Approach*

Collaborative filtering (CF) employing a consumer preference database to make personal product recommendations is achieving widespread success in E-commerce. However, it does not scale well to the ever-growing number of consumers. The quality of the recommendation also needs to be improved in order to gain more trust from consumers. This paper attempts to improve the accuracy and efficiency of collaborative filtering. We present a unified information-theoretic approach to measure the relevance of features and instances. Feature weighting and instance selection methods are proposed for collaborative filtering. The proposed methods are evaluated on the well-known EachMovie data set and the experimental results demonstrate a significant improvement in accuracy and efficiency.*This work was performed in Corporate Technology, Siemens AG.     

Student perceptions of collaborative learning,social presence and satisfaction in a blended learning environment: Relationships and critical factors

The purpose of this study was to examine the relationships of the students’ perceived levels of collaborative learning, social presence and overall satisfaction in a blended learning environment. This research studied the relationship of these three variables and identified critical factors related to them. The participants were 48 graduate students who took a blended-format course in health education and worked on a collaborative group project related to the development of a comprehensive HIV-AIDS prevention plan. Data was collected from the Student Perception Questionnaire and face-to-face interviews. The analysis of quantitative data indicated that student perceptions of collaborative learning have statistically positive relationships with perceptions of social presence and satisfaction. This means that students who perceived high levels of collaborative learning tended to be more satisfied with their distance course than those who perceived low levels of collaborative learning. Similarly, students with high perceptions of collaborative learning perceived high levels of social presence as well. Surprisingly, the relationship between social presence and overall satisfaction was positive but not statistically significant. Interview data revealed that (a) course structure, (b) emotional support, and (c) communication medium were critical factors associated with student perceptions of collaborative learning, social presence, and satisfaction. Explanations about findings and implications for instructional design are discussed in the conclusion.     

Discovering user communities on the Internet using unsupervised machine learning techniques

Interest in the analysis of user behaviour on the Internet has been increasing rapidly, especially since the advent of electronic commerce. In this context, we argue here for the usefulness of constructing communities of users with common behaviour, making use of machine learning techniques. In particular, we assume that the users of any service on the Internet constitute a large community and we aim to construct smaller communities of users with common characteristics. The paper presents the results of three case studies for three different types of Internet service: a digital library, an information broker and a Web site. Particular attention is paid on the different types of information access involved in the three case studies: query-based information retrieval, profile-based information filtering and Web-site navigation. Each type of access imposes different constraints on the representation of the learning task. Two different unsupervised learning methods are evaluated: conceptual clustering and cluster mining. One of our main concerns is the construction of meaningful communities that can be used for improving information access on the Internet. Analysis of the results in the three case studies brings to surface some of the important properties of the task, suggesting the feasibility of a common methodology for the three different types of information access on the Internet.     

Active learning through collaborative knowledge building using an automatic free-text scoring system in a b-learning environment

According to active learning, students should be responsible for their own learning. Automatic free-text scoring allows teachers to provide open-ended questions with their correct answers to a computer system, so when students answer the questions, they get immediate feedback (a score, a comment, or both). However, teachers are usually overloaded with many tasks, and they may not have time to create the questions with the correct answers. Therefore, in the 2012/2013 academic year, we asked a group of 124 Pre-Primary and Primary Education students to become the creators of the questions and their correct answers in groups in a free-text scoring system, so the questions use learners’ language, not teachers’ language. From them, 41 students (group of involved students, GIS) fulfilled all the requirements during the course. Our hypothesis was that GIS would be able to increase their academic performance and levels of engagement compared to the rest of the students. The results gathered provide statistic evidence to support that hypothesis. This study pretends to help teachers who want to increase the academic performance and levels of engagement of their students in courses that they may find boring and unrelated to the main topic of their degree, or not directly related with their main academic interests.     

Enhancing empathy instruction using a collaborative virtual learning environment for children with autistic spectrum conditions

Empathy is an essential part of normal social function that people with autism spectrum conditions (ASCs) lack. This study uses the intervention of enhancing empathy via 3D animated scenarios of empathy in a virtual learning environment to help those deficient in empathy. Specifically, this study explores the understanding of empathy, perspective-taking and the performance of understanding of empathy via a collaborative virtual learning environment (CVLE) - empathy system. The study, which used CVLE - 3D empathy systems and three participants diagnosed with ASCs, conducted multiple baseline research for evidence of improved understanding of empathy via system usage. This experimental study lasted 5 months and the experimental results indicate that using the CVLE 3D empathy system had significant and positive effects on participant use of empathy, both within the CVLE 3D empathy system and in terms of maintaining learning in understanding empathy.