Search

Graph clustering is a fundamental task to discover community ties in multi-layer networks. In this paper, we propose a network embedding technique to find covert communities in multi-layer dark networks using a Log-BiLinear (LBL) approach. Recent works on graph clustering using network embedding have focused on new ways of learning representations of nodes and relations, upon which a classic clustering method is then used to identify the communities (clusters). However, these embedding approach does not yield good and accurate communities from the clustering task. Hence, we address this issue with a sequence-based network embedding technique on a multi-layer network. Our proposal learns structural representations of nodes and relations simultaneously by capturing the position of a given node within a set of neighboring anchor-set, and the type of connections between nodes in the anchor-set. To find the clusters (communities), clustering centroids are also learned as the representations of nodes and relations are extracted. Our solution is well-suited to detecting covert communities, such as terrorist networks. In our experiments on three real-world terrorist datasets and one synthetic network, our approach is found to deliver a higher level of accuracy in detecting covert communities compared with six baseline methods.

Conventional microtask crowdsourcing platforms rely on a random task distribution strategy and repeatedly assign tasks to workers. This strategy known as repeated labelling suffers from two shortcomings of high cost and low accuracy as a result of making random distributions. To overcome such shortcomings researchers have introduced task assignment as a substitute strategy. In this strategy, an algorithm selectively chooses suitable tasks for an online worker. Hence, task assignment has gained attentions from researchers to reduce the cost of microtasking whiling increasing its accuracy. However, the existing algorithms on task assignment suffer from four shortcomings as: (i) human intervention, (ii) reliance on a rough estimation of ground truth, (iii) reliance on workers’ dynamic capabilities and (iv) lack of ability in dealing with sparsity. To overcome these shortcomings this paper proposes a new task assignment algorithm known as LEarning Automata based Task assignment (LEATask), that works based on the similarities of workers in performance. This algorithm has two stages of exploration and exploitation. In exploration stage, first a number of workers are hired to learn their reliability. Then, LEATask clusters the hired workers using a given clustering algorithm, and for each cluster generates learning automata. Later, the clusters of workers along with their attached learning automata will be used in exploitation stage. Exploitation stage initially assigns a number of tasks to a newly arrived worker to learn the worker’s reliability. Then, LEATask identifies the cluster of worker. Based on the cluster that worker resides in and the attached learning automata, the next tasks will be assigned to the new worker. LEATask has been empirically evaluated using several real datasets and compared against the baseline and novel algorithms, in terms of root mean square error. The comparisons indicates LEATask consistently is showing better or comparable performance.

Timely identification of terrorist networks within civilian populations could assist security and intelligence personnel to disrupt and dismantle potential terrorist activities. Finding “small” and “good” communities in multi-layer terrorist networks, where each layer represents a particular type of relationship between network actors, is a vital step in such disruption efforts. We propose a community detection algorithm that draws on the principles of discrete-time random walks to find such “small” and “good” communities in a multi-layer terrorist network. Our algorithm uses several parallel walkers that take short independent random walks towards hubs on a multi-layer network to capture its structure. We first evaluate the correlation between nodes using the extracted walks. Then, we apply an agglomerative clustering procedure to maximize the asymptotical Surprise, which allows us to go beyond the resolution limit and find small and less sparse communities in multi-layer networks. This process affords us a focused investigation on the more important seeds over random actors within the network. We tested our algorithm on three real-world multi-layer dark networks and compared the results against those found by applying two existing approaches – Louvain and InfoMap – to the same networks. The comparative analysis shows that our algorithm outperforms the existing approaches in differentiating “small” and “good” communities.

This paper presents a classification framework and a systematic analysis of literature on answer aggregation techniques for the most popular and important type of crowdsourcing, i.e., micro-task crowdsourcing. In doing so, we analyzed research articles since 2006 and developed four classification taxonomies. First, we provided a classification framework based on the algorithmic characteristics of answer aggregation techniques. Second, we outlined the statistical and probabilistic foundations used by different types of algorithms and micro-tasks. Third, we provided a matrix catalog of the data characteristics for which an answer aggregation algorithm is designed. Fourth, a matrix catalog of the commonly used evaluation metrics for each type of micro-task was presented. This paper represents the first systematic literature analysis and classification of the answer aggregation techniques for micro-task crowdsourcing.

Contradictory results are documented in the literature regarding which type of mutual fund has superior performance; an Islamic or conventional mutual fund. Due to the relative short history of the Islamic mutual funds' industry, prior literature has inevitably relied on a small sample size with a short sample period. With the longest applicable sample period, this study represents one of the most recent attempts to address this conflicting evidence. We find there is no clear cut over performance by Islamic mutual funds against their conventional peers across the three financial crises in our sample period, with the exception of the most recent global financial crisis, where Islamic mutual funds generally outperformed their conventional counterparts. We further find that Islamic funds significantly outperformed conventional funds in the riskiest asset class, equity, one year before and during the global financial crisis. We further reveal that the modified value at risk for Islamic mutual funds was significantly lower than their conventional peers during the global financial crisis. This seems to indicate that Islamic mutual funds have better risk management compared to conventional peers.

Misclassification costs of minority class data in real-world applications can be very high. This is a challenging problem especially when the data is also high in dimensionality because of the increase in overfitting and lower model interpretability. Feature selection is recently a popular way to address this problem by identifying features that best predict a minority class. This paper introduces a novel feature selection method call SYMON which uses symmetrical uncertainty and harmony search. Unlike existing methods, SYMON uses symmetrical uncertainty to weigh features with respect to their dependency to class labels. This helps to identify powerful features in retrieving the least frequent class labels. SYMON also uses harmony search to formulate the feature selection phase as an optimisation problem to select the best possible combination of features. The proposed algorithm is able to deal with situations where a set of features have the same weight, by incorporating two vector tuning operations embedded in the harmony search process. In this paper, SYMON is compared against various benchmark feature selection algorithms that were developed to address the same issue. Our empirical evaluation on different micro-array data sets using G-Mean and AUC measures confirm that SYMON is a comparable or a better solution to current benchmarks.

There are now a growing number of social networking websites with millions of users, creating a fertile ground for “spammers” to abuse opportunities in these websites for their own gain through constant exposure of malicious communications to other users. The variety of interactions afforded by these social networks has resulted in a Multiplex Network of interactions. In these networks, malicious users evade detection by frequently changing the nature of their activities. This makes it challenging to analyse users’ interactions to capture anomalous behaviours. In this paper, we aimed to detect spammers in a large time-evolving multiplex social network called Tagged.com. For this purpose, we used four different sets of features: (i) a set of light-weight behavioural features to capture the structural behaviour of users in their neighbourhood network; (ii) a set of bursty features and (iii) sequence-based features for capturing the temporal behaviour of users; and (iv) a set of profile-based features which was used as a side information. In addition, we also employed an unsupervised Laplacian Score based approach for feature selection and space dimensionality reduction. The experimental results showed an accuracy of over 88% in spammer detection with a lower empirical time complexity for feature extraction. Implementing behavioural and bursty features in a relational data management system makes the proposed approach more practical since most of the real-world networks store their data in relational databases.

Time control plays a critical role within the online mastery learning (OML) approach. This paper examines the two commonly implemented mastery learning strategies – personalised system of instructions and learning for mastery (LFM) – by focusing on what occurs when there is an instructional time constraint. Using a large data set from a postgraduate finance course offered at an Australian university, we explore students' online quiz-completion patterns, then empirically investigate whether the imposition of an instructional time constraint in the OML approach has an impact on their final-examination performance. Our results suggest that the LFM strategy with an instructional time constraint has a positive impact on students' learning behaviour and contributes to better overall academic performance. Further, our findings suggest that facilitators should be encouraged to implement an instructional time constraint when adopting an OML approach.

Community detection (CD) is the act of grouping similar objects. This has applications in social networks. The conventional CD algorithms focus on finding communities from one single perspective (objective) such as structure. However, reliance on only one objective of structure. This makes the algorithm biased, in the sense that objects are well separated in terms of structure, while weakly separated in terms of other objective function (e.g., attribute). To overcome this issue, novel multi-objective community detection algorithms focus on two objective functions, and try to find a proper balance between these two objective functions. In this paper we use Harmony Search (HS) algorithm and integrate it with Pareto Envelope-Based Selection Algorithm 2 (PESA-II) algorithm to introduce a new multi-objective harmony search based community detection algorithm. The integration of PESA-II and HS helps to identify those non-dominated individuals, and using that individuals during improvisation steps new harmony vectors will be generated. In this paper we experimentally show the performance of the proposed algorithm and compare it against two other multi-objective evolutionary based community detection algorithms, in terms of structure (modularity) and attribute (homogeneity). The experimental results indicate that the proposed algorithm is outperforming or showing comparable performances.

Purpose Ever since the publication of the original Jegadeesh and Titman (1993) study, momentum effect has been tested vigorously to validate its pervasiveness for different time periods and across different markets. In spite of numerous out-of-sample tests, there is one apparent alibi – little research has been devised for steady increasing of Shari’ah compliant stocks. Methodology/approach This study is to examine the momentum strategy returns in a global Shari’ah compliant stock setting. Findings It finds strong presence of stock momentum returns for Pakistan and Malaysia. And the momentum returns are neither driven by industry momentum nor by the small size stocks. Though no momentum profits are found for the portfolios formed by global Shari’ah compliant stocks, this seems to be largely due to return reversal for the small size Shari’ah compliant stocks. Originality/value The strong presence of momentum profits for relatively large Shari’ah compliant stocks is a desirable trait as it indicates that the momentum trading strategies are practical and implementable.

Estimation of worker reliability on microtask crowdsourcing platforms has gained attention from many researchers. On microtask platforms no worker is fully reliable for a task and it is likely that some workers are spammers, in the sense that they provide a random answer to collect the financial reward. Existence of spammers is harmful as they increase the cost of microtasking and will negatively affect the answer aggregation process. Hence, to discriminate spammers and non-spammers one needs to measure worker reliability to predict how likely that a worker put an effort in solving a task. In this paper we introduce a new reliability estimation algorithm works based on bee colony algorithm called REBECO. This algorithm relies on Gaussian process model to estimate reliability of workers dynamically. With bees that go in search of pollen, some are more successful than the others. This maps well to our problem, where some workers (i.e., bees) are more successful than other workers for a given task thus, giving rise to a reliability measure. Answer aggregation with respect to worker reliability rates has been considered as a suitable replacement for conventional majority voting. We compared REBECO with majority voting using two real world datasets. The results indicate that REBECO is able to outperform MV significantly.

Crowdsourcing refers to distributing microtasks to an unknown group of online workers. Given that workers have varying expertise levels, a major research challenge for crowdsourcing is solving the problems of untargeted task assignment and unestimated aggregation of results. Although existing approaches can estimate the expertise of workers and use expertise information to allocate tasks, the effectiveness of these approaches is limited for the following reasons: 1) reliance on human intervention; 2) dependence on the type of answers; 3) non-sparseness; 4) post-expertise estimation. To overcome these limitations of crowdsourcing, this paper introduces an unsupervised approach to expertise estimation in microtask crowdsourcing that is independent of answer type, which is named ROUgh set based eXpertise estimation (ROUX). We consider the problem of expertise estimation as a metaheuristic optimization search problem, and integrate it with a rough set to better estimate the expertise of each online worker. Further, ROUX uses the expertise rating of workers for task assignment to maximize the accuracy of the results and lower the cost. Extensive experimental evaluations using real-world datasets show that ROUX performs remarkably in accuracy improvement and cost efficacy.

It is widely accepted that some firms' attributes or characteristics, such as a firm's size or book-to-market ratio, attract premiums in terms of average returns, which is a pervasive phenomenon not restricted to just a few individual markets. However, the way to derive these premiums by sorting firms based on their characteristics that are known to be associated with share returns, is not without controversy. This study takes a totally different approach by adopting a novel Self-Organizing Maps approach to cluster share returns first and identify the attributable factors afterwards. It finds eminent presence of the value effect in the Chinese A-shares market. It also finds that two other firm attributes, ROA and cash-flow-to-price, also have explanatory power over share returns. Liquidity and the ratio of tradable to total shares, which are often used to proxy corporate governance, show little evidence of explaining share return dispersion. Surprisingly, a “reversed” size effect is reported in this study.

The capability of identifying customers who are more likely to respond to a product is an important issue in direct marketing. This paper investigates the impact of feature selection on predictive models which predict reordering demand of small and medium-sized enterprise customers in a large online job-advertising company. Three well-known feature subset selection techniques in data mining, namely correlation-based feature selection (CFS), subset consistency (SC) and symmetrical uncertainty (SU), are applied in this study. The results show that the predictive models using SU outperform those without feature selection and those with the CFS and SC feature subset evaluators. This study has examined and demonstrated the significance of applying the feature-selection approach to enhance the accuracy of predictive modelling in a direct-marketing context.

Research to solve engineering and science problems commonly require the collection and complex analysis of a vast amount of data. This makes them a natural exemplar of big data applications. For example, data from weather stations, high resolution images from CT scans, or data captured by astronomical instruments all easily showcase one or more big data characteristics, i.e., volume, velocity, variety and veracity. These big data characteristics present computational and analytical challenges that need to be overcame in order to deliver engineering solutions or make scientific discoveries. In this chapter, we catalogued engineering and science problems that carry a big data angle. We will also discuss the research advances for these problems and present a list of tools available to the practitioner. A number of big data application exemplars from the past works of the authors are discussed with further depth, highlighting the association of the specific problem and its big data characteristics. The overview from these various perspectives will provide the reader an up-to-date audit of big data developments in engineering and science.

Business intelligence and analytics (BIA) initiatives are costly, complex and experience high failure rates. Organizations require effective approaches to evaluate their BIA capabilities in order to develop strategies for their evolution. In this paper, we employ a design science paradigm to develop a comprehensive BIA effectiveness diagnostic (BIAED) framework that can be easily operationalized. We propose that a useful BIAED framework must assess the correct factors, should be deployed in the proper process context and acquire the appropriate input from different constituencies within an organization. Drawing on the BIAED framework, we further develop an online diagnostic toolkit that includes a comprehensive survey instrument. We subsequently deploy the diagnostic mechanism within three large organizations in North America (involving over 1500 participants) and use the results to inform BIA strategy formulation. Feedback from participating organizations indicates that BIA diagnostic toolkit provides insights that are essential inputs to strategy development. This work addresses a significant research gap in the area of BIA effectiveness assessment.

Musical preference has long been a research interest in the field of music education, and studies consistently confirm the importance of musical preference in one's musical learning experiences. However, only a limited number of studies have been focussed on the field of early childhood education (e.g., Hargreaves, North, & Tarrant, 2006; Roulston, 2006). Further, among these limited early childhood studies, few of them discuss children's musical preference in both the East and the West. There is very limited literature (e.g., Faulkner et al., 2010; Szymanska, 2012) which explores the data by using a data mining approach. This study aims to bridge the research gaps by examining children's musical preference in Hong Kong and in South Australia by applying a data mining technique - Self Organizing Maps (SOM), which is a clustering method that groups similar data objects together. The application of SOM is new in the field of early childhood education and also in the study of children's musical preference. This paper specifically aims to expand a previous study (Yim & Ebbeck, 2009) by conducting deeper investigations into the existing datasets, for the purpose of uncovering insights that have not been identified through data mining approach. Refereed/Peer-reviewed

It is widely accepted that the presence of some of the firm’s attributes or characteristics attracting premiums in terms of average returns is pervasive and not restricted to a few individual markets. However, the way to derive these premiums by sorting firms based on their characteristics that are known associated with share returns is not without controversy. This chapter takes a different approach by adopting a novel Multi Self-Organising Maps to cluster shares first and then identify fundamental factors afterwards. It finds that firm’s size and book-to-market ratio attributes do have explanatory power over share average returns. There is also lack of evidence for other factors in explaining the share average returns.

With the massive amount of crime data generated daily, this has put law enforcement under intensive stress. This means that law enforcement has to compete against the time to solve crime. In addition, the focus of crime investigation has been expanded from the ability to catch the criminals towards the ability to act before a crime happens (i.e pre-crime). Given such situation, creation of crime profiles is very important to law enforcement, especially in understanding the behaviours of criminals and identifying the characteristics of similar crimes. In fact, crime profiles could be used to solve similar crimes and thus pre-crime action could be conducted. In this paper, a brain inspired conceptual model is proposed and a structurally adaptive neural network is deployed for its implementation. Subsequently, the proposed model is applied for the identification and presentation of multi-view crime patterns. Such multi-view crime patterns could be useful for the construction of crime profiles. Moreover, the suitability of the proposed model in crime profiling is discussed and demonstrated through some experimental results.

The appearance of patterns could be found in different modalities of a domain, where the different modalities refer to the data sources that constitute different aspects of a domain. Particularly, the domain of our discussion refers to crime and the different modalities refer to the different data sources such as offender data, weapon data, etc. in crime domain. In addition, patterns also exist in different levels of granularity for each modality. In order to have a thorough understanding a domain, it is important to reveal the hidden patterns through the data explorations at different levels of granularity and for each modality. Therefore, this paper presents a new model for identifying patterns that exist in different levels of granularity for different modes of crime data. A hierarchical clustering approach - growing self organising maps (GSOM) has been deployed. Furthermore, the model is enhanced with experiments that exhibit the significance of exploring data at different granularities.

Biologically human brain processes information in both unimodal and multimodal approaches. In fact, information is progressively abstracted and seamlessly fused. Subsequently, the fusion of multimodal inputs allows a holistic understanding of a problem. The proliferation of technology has exponentially produced various sources of data, which could be likened to being the state of multimodality in human brain. Therefore, this is an inspiration to develop a methodology for exploring multimodal data and further identifying multi-view patterns. Specifically, we propose a brain inspired conceptual model that allows exploration and identification of patterns at different levels of granularity, different types of hierarchies and different types of modalities. A structurally adaptive neural network is deployed to implement the proposed model. Furthermore, the acquisition of multi-view patterns with the proposed model is demonstrated and discussed with some experimental results., {"references":["C. B. Saper, S. Iversen, and R. Frackowiak, \"Integration of sensory\nand motor function: The association areas of the cerebral cortex and\nthe cognitive capabilities of the brain,\" in Principles of neural science,\n4th ed., E. R. Kandel, J. H. Schwartz, and T. M. Jessell, Eds. New\nYork: McGraw-Hill, 2000, ch. 19, pp. 349-380.","V. Mountcastle, \"The columnar organization of the neocortex.\" Brain,\nno. 120, pp. pp. 701-722, 1997.","K. Friston, \"Hierarchical models in the brain,\" PLoS Computational\nBiology, vol. 4, no. 11, 2008.","D. L. Hall and J. Llinas, \"An introduction to multisensor data fusion,\"\nProceedings of the IEEE, vol. 85, no. 1, pp. pp. 6-23, 1997.","J. R. Hobbs, \"Granularity,\" in Proceedings of the 9th International Joint\nConference on Artificial Intelligence (IJCAI). Los Angeles, USA:\nMorgan Kaufmann, 1985, pp. 432-435.","Y. Yao, \"Perspectives of granular computing,\" in IEEE International\nConference on Granular Computing (GrC), Beijing, China, 2005, pp.\n85-90.","Y. Chen and Y. Yao, \"A multiview approach for intelligent data analysis\nbased on data operators,\" International Journal of Information Sciences,\nvol. 178, pp. pp. 1-20, 2008.","ÔÇöÔÇö, \"Multiview intelligent data analysis based on granular computing,\"\nin IEEE International Conference on Granular Computing (GrC),\nAtlanta, USA, 2006, pp. 281-286.","M. Minsky, The Emotion Machine : commensense thinking, artificial\nintelligence, and the future of the human mind. New York: Simon &\nSchuster, 2006.\n[10] S. Zhang, C. Zhang, and X. Wu, Knowledge Discovery in Multiple\nDatabases. London, UK: Springer-Verlag, 2004.\n[11] N. Kasabov, E. Postma, and J. van den Herik, \"Avis: a connectionistbased\nframework for integrated auditory and visual information processing,\"\nInformation Sciences, vol. 123, pp. pp. 127-148, 2000.\n[12] N. Kasabov, \"Evolving systems for integrated multi-modal information\nprocessing,\" in Evolving Connectionist Systems: Methods and Applications\nin Bioinformatics, Brain Study and Intelligent Machine. London:\nSpringer-Verlag, 2003, ch. 13, pp. 257-271.\n[13] ÔÇöÔÇö, \"Evolving intelligent systems for adaptive multimodal information\nprocessing,\" in Evolving Connectionist Systems The Knowledge Engineering\nApproach. London: Springer-Verlag, 2007, ch. 13, pp. 361-\n380.\n[14] J. Hawkins and S. Blakeslee, On Intelligence. New York: Times Books,\n2004.\n[15] J. Hawkins and D. George, \"Hierarchical temporal memory - concepts,\ntheory, and terminology,\" Numenta Inc., Redwood City, California,\nWhite Paper, 2007.\n[16] Y. Lu, \"Concept hierarchy in data mining: Specification,generation and\nimplementation,\" Master-s thesis, Simon Fraser University, Canada,\n1997.\n[17] D. Alahakoon, S. K. Halgamuge, and B. Srinivasan, \"A self growing\ncluster development approach to data mining,\" in IEEE Conference\nSystems, Man and Cybernetics, San Diego, USA, 1998, pp. 2901-2906.\n[18] ÔÇöÔÇö, \"Dynamic self-organizing maps with controlled growth for knowledge\ndiscovery,\" IEEE Transactions on Neural Networks, vol. 11, no. 3,\npp. pp. 601-614, 2000.\n[19] R. S. Forsyth, \"Uci machine learning repository, zoo data set,\" 1990.\n(Online). Available: http://archive.ics.uci.edu/ml/datasets/Zoo"]}