Using trust for secure collaboration in uncertain environments

The SECURE project investigates the design of security mechanisms for pervasive computing based on trust. It addresses how entities in unfamiliar pervasive computing environments can overcome initial suspicion to provide secure collaboration.     

Exploiting data mining techniques for broadcasting data in mobile computing environments

Mobile computers can be equipped with wireless communication devices that enable users to access data services from any location. In wireless communication, the server-to-client (downlink) communication bandwidth is much higher than the client-to-server (uplink) communication bandwidth. This asymmetry makes the dissemination of data to client machines a desirable approach. However, dissemination of data by broadcasting may induce high access latency in case the number of broadcast data items is large. We propose two methods aiming to reduce client access latency of broadcast data. Our methods are based on analyzing the broadcast history (i.e., the chronological sequence of items that have been requested by clients) using data mining techniques. With the first method, the data items in the broadcast disk are organized in such a way that the items requested subsequently are placed close to each other. The second method focuses on improving the cache hit ratio to be able to decrease the access latency. It enables clients to prefetch the data from the broadcast disk based on the rules extracted from previous data request patterns. The proposed methods are implemented on a Web log to estimate their effectiveness. It is shown through performance experiments that the proposed rule-based methods are effective in improving the system performance in terms of the average latency as well as the cache hit ratio of mobile clients.     

多MapReduce作业协同下的大数据挖掘类算法资源效率优化

由于任意的MapReduce作业都需要独立地进行任务调度、资源分配等一系列复杂的操作,这使得同一算法协同的多个MapReduce作业之间,存在着大量的冗余磁盘I/O及资源重复申请操作,导致计算过程中资源利用效率低下。大数据挖掘类算法通常被切分成多个MapReduce job协作完成。以ItemBased算法为例,对多MapReduce作业协同下的大数据挖掘算法存在的资源效率问题进行了分析,提出基于DistributedCache的ItemBased算法,利用DistributedCache将多个MapReduce job之间的I/O数据进行缓存处理,打破作业之间独立性的缺陷,减少map与reduce任务之间的等待时延。实验结果表明,DistributedCache能够提高MapReduce作业的数据读取速度,利用DistributedCache重构后的算法极大地减少了map与reduce任务之间的等待时延,资源效率提高3倍以上。     

Using a grounded theory approach to study online collaboration behaviors

This paper discusses how an interpretive theory of action was explored and developed through iterative cycles of grounded theory generation. We establish our motivation for employing the grounded theory method in an area that is overflowing with theories of learning, then move on to the practicalities of generating an interpretive grounded theory by following the ‘vapor trails’ left by online learners. We describe how we incorporated the use of mixed methods into an interpretive grounded theory process, with a theoretical sampling strategy that used ‘complementary comparison’ to feed back into a new cycle of constant comparison. We discuss how constant comparison may be enhanced by researcher debate around emerging themes and categories, co-coding of data samples, coding of researcher theoretical memos, and reflection-in-action during explicit explanations of coding schemes to research assistants and the review of research process memos. Finally, we discuss how and why the substantive theory of action that was generated by this process provides an original contribution to theories of collaborative online learning by accounting for both visible and invisible learning strategies that explain the role of thought-leaders in a community of inquiry and account for vicarious learning.     

Using signature techniques for information filtering in wireless and mobile environments

This paper discusses the issue of power conservation on mobile clients, e.g., palmtop, in wireless and mobile environments. It suggests that techniques using signatures are suitable for realtime information filtering on mobile clients. Three signature-based approaches, namely simple signature, integrated signature and multi-level signature schemes, are presented. The cost models for the access time and tune-in time of these three approaches are developed. We show that the multi-level signature method is in general better than the other two methods.Recommended by: Daniel Barbara, Ravi Jain and Narayanan Krishnakumar     

面向大数据流式计算的任务管理技术综述

流式计算是大数据的一种重要计算模式,大数据流式计算已成为研究热点。任务管理是大数据流式计算的核心功能之一,负责对流式计算的任务进行资源调度及全生命周期管理。目前对于大数据流式计算的技术调研工作主要集中于流式计算应用需求、体系结构及整体技术,缺乏对大数据流式计算任务管理技术的精细化调研分析。首先给出流式计算任务管理的抽象功能模型,其次基于该模型对任务管理的关键技术进行了分类和综述,最后对既有主流的大数据流式计算系统对上述关键技术的应用、集成和优化进行了调研分析。     

Cohesion in online environments

International Journal of Computer-Supported Collaborative Learning - This paper presents a study of group cohesion as it arises in online small group different time and place collaboration....     

Social influence determination on big data streams in an online social network

Social networks have become a good place to promote products and also to campaign for causes. Maximizing the spread of information in an online social network at a least cost has attracted the attention of publicist’s. In general, influence user ranking methods are derived either by a network’s topological features or by user features but not both. Existing Influence Maximization Problem (IMP) operates as a modification of greedy algorithms that cannot scale streaming data. Which are time consuming and cannot handle large networks because it requires heavy Monte-Carlo simulation. This is also an NP hard problem in both linear threshold and independent cascade models. Our proposed work aims to address IMP through a Rank-based sampling approach in the Map-Reduce environment. This novel technique combines user and topological features of the network enabling it to handle real-time streaming data. Our experiment of influenced rank-based sampling approach to influence maximization is compared to the greedy approach with and without sampling that exhibits an accuracy of 82%. Performance analysis in terms of running time is reduced from O(n ³) to O(k n). Where ‘k’ is the size of the sample dataset and ‘n’ is the number of user’s.     

Using Information Retrieval techniques for supporting data mining

The classic two-stepped approach of the Apriori algorithm and its descendants, which consisted of finding all large itemsets and then using these itemsets to generate all association rules has worked well for certain categories of data. Nevertheless for many other data types this approach shows highly degraded performance and proves rather inefficient.

We argue that we need to search all the search space of candidate itemsets but rather let the database unveil its secrets as the customers use it. We propose a system that does not merely scan all possible combinations of the itemsets, but rather acts like a search engine specifically implemented for making recommendations to the customers using techniques borrowed from Information Retrieval.     

Using online collaboration applications for group assignments: The interplay between design and human characteristics

In recent years, educators and students are increasingly employing online collaboration applications such as Google Docs™ and PBWorks™ for group projects and assignments. Yet, the effectiveness of these emerging technologies has not been rigorously examined. Anchoring upon and informed by the existing literature, two design characteristics – sociability and visibility, and two human characteristics – gender and age, are focused on, which are salient in online collaboration applications. A field experiment was conducted to examine the direct and moderating effects of design and human characteristics on learning outcomes. The research found that sociability improved process satisfaction and positive social environment while visibility enhanced academic performance and solution satisfaction of learners. Males had higher solution satisfaction while older learners had higher academic performance. Moderating effects were also found. Both theoretical and practical implications are drawn. In particular, a rubric for online collaboration application selection for academic performance is conceived. This study provides empirical support for online collaboration application effectiveness in education which will augur well for future adoption, use and evolution.     

一种基于遗传算法优化的大数据特征选择方法

提出了一种基于遗传算法的大数据特征选择算法。该算法首先对各维度的特征进行评估,根据每个特征在同类最近邻和异类最近邻上的差异度调整其权重,基于特征权重引导遗传算法的搜索,以提升算法的搜索能力和获取特征的准确性;然后结合特征权重计算特征的适应度,以适应度作为评价指标,启动遗传算法获取最优的特征子集,并最终实现高效准确的大数据特征选择。通过实验分析发现,该算法能够有效减小分类特征数,并提升特征分类准确率。     

Industry classification with online resume big data: A design science approach

Industry classification is a vital step of industry analysis and competitive intelligence. However, existing schemes and methods are limited by the lagged information of firms’ business and the lack of consideration of the human resource aspects. In this paper, we adopt a design science approach to develop and evaluate a novel industry classification method by constructing a labor mobility network using online resume big data collected from the professional social network. We also propose a hierarchical extension of the community detection algorithm to better discover scalable firm clusters on the constructed network. The evaluation conducted on real-world datasets shows that our method outperforms the existing industry classification schemes and the state-of-the-art methods by improving their explanatory power and enlarging the cross-industry variation. Moreover, two application cases confirm the validity of our method in earlier revealing firms’ action of entering new industries.     

Text analytics for big data using rough–fuzzy soft computing techniques

Text mining or analytics is important for various applications such as market analysis and biomedical purposes because it enables the efficient retrieval of information from large datasets. During the analysis, increasing the dimensionality of the data reduces the performance of an entire system because doing so may retrieve irrelevant text, which creates errors. Therefore, this paper introduces big data and data mining techniques to analyse large volumes of information while mining texts, emails, blogs, online forums, news, and call centre documents. Initially, the data are collected from various sources that contain noise, which is removed by applying normalization techniques. Data mining techniques eliminate the irrelevant information and noise, and the relevant features are selected using the rough set‐based particle swarm optimization algorithm. The selected features are formed as a cluster using a fuzzy set with the particle swarm optimization algorithm, which improves the efficiency of the mining process. Then, the efficiency of the system is evaluated using the University of California Irvine Machine Learning Repository knowledge process mining database, along with the sum of the intra cluster distances, the mean squared error rate, and the accuracy.     

Open media service architecture for advanced collaboration environments

Advanced collaboration environments are extensively utilized for distance learning, e-science, and other distributed global collaboration events. In such environments, high-quality and seamless media services play an important role in improving the quality of user experience to participants. In this paper, to support high-quality media-based services, we design open media service architecture for advanced collaboration environments, by combining the open interface for state-of-the-art media tools, the performance monitoring tools for devices and networks, and application-level adaptation schemes for media streaming. By implementing the proposed architecture on top of an open-source Access Grid (AG) collaboration toolkit, we verify that high-quality collaboration among several collaboration sites can be effectively realized over a multicast-enabled network testbed with improved media quality experience.

Enabling OLAP in mobile environments via intelligent data cube compression techniques

The main drawbacks of handheld devices (small storage space, small size of the display screen, discontinuance of the connection to the WLAN etc) are often incompatible with the need of querying and browsing information extracted from enormous amounts of data which are accessible through the network. In this application scenario, data compression and summarization have a leading role: data in a lossy compressed format can be transmitted more efficiently than the original ones, and can be effectively stored in handheld devices (setting the compression ratio accordingly). In this paper, we introduce a very effective compression technique for multidimensional data cubes, and the system Hand-OLAP, which exploits this technique to allow handheld devices to extract and browse compressed two-dimensional OLAP views coming from multidimensional data cubes stored on a remote OLAP server localized on the wired network. Hand-OLAP effectively and efficiently enables OLAP in mobile environments, and also enlarges the potentialities of Decision Support Systems by taking advantage from the “naturally” decentralized nature of such environments. The idea which the system is based on is: rather than querying the original multidimensional data cubes, it may be more convenient to generate a compressed OLAP view of them, store such view into the handheld device, and query it locally (off-line), thus obtaining approximate answers that are suitable for OLAP applications.     

Applying machine learning techniques for scaling out data quality algorithms in cloud computing environments

Deduplication is the task of identifying the entities in a data set which refer to the same real world object. Over the last decades, this problem has been largely investigated and many techniques have been proposed to improve the efficiency and effectiveness of the deduplication algorithms. As data sets become larger, such algorithms may generate critical bottlenecks regarding memory usage and execution time. In this context, cloud computing environments have been used for scaling out data quality algorithms. In this paper, we investigate the efficacy of different machine learning techniques for scaling out virtual clusters for the execution of deduplication algorithms under predefined time restrictions. We also propose specific heuristics (Best Performing Allocation, Probabilistic Best Performing Allocation, Tunable Allocation, Adaptive Allocation and Sliced Training Data) which, together with the machine learning techniques, are able to tune the virtual cluster estimations as demands fluctuate over time. The experiments we have carried out using multiple scale data sets have provided many insights regarding the adequacy of the considered machine learning algorithms and proposed heuristics for tackling cloud computing provisioning.     

Context-dependent awareness support in open collaboration environments

The widespread adoption of online services for performing work, home and leisure tasks enables users to operate in the ubiquitous environment provided by the Internet by managing a possibly high number of parallel (private and shared) activity contexts. The provision of awareness information is a key factor for keeping users up-to-date with what happens around them; e.g., with the operations performed by their collaborators. However, the delivery of notifications describing the occurred events can interrupt the users’ activities, with a possible disruptive effect on their emotional and attentional states. As a possible solution to the trade-off between informing and interrupting users, we defined two context-dependent notification management policies which support the selection of the notifications to be delivered on the basis of the user’s current activities, at different granularity levels: general collaboration context versus task carried out. These policies are offered by the COntext depeNdent awaReness informAtion Delivery (CONRAD) framework. We tested such policies with users by applying them in a collaboration environment that includes a set of largely used Web 2.0 services. The experiments show that our policies reduce the levels of workload on users while supporting an up-to-the-moment understanding of the interaction with their shared contexts. The present paper presents the CONRAD framework and the techniques underlying the proposed notification policies.