I received my MSc in Computer Science from Wroclaw University of Science and Technology (WrUST), Poland, and an MSc in Software Engineering from the Blekinge Institute of Technology, Sweden, both in 2006. I was awarded my PhD in Computer Science in November 2009 from Wroclaw University of Science and Technology, and in the same year I was appointed a Senior Visiting Research Fellow at Bournemouth University (BU), where from 2010 I was a Lecturer in Informatics. I joined King’s in November 2011 as a Lecturer in Computer Science. In September 2015 I returned to Bournemouth University as a Principal Academic in Computing where I was also a Head of SMART Technology Research Group and a member of Data Science Initiative. In September 2017 I joined UTS as Associate Professor in Network Science.
Over the last few years I have collaborated with various commercial organizations and research groups. These research efforts resulted in publication of over 80 research papers in journals, books and conference proceedings. I have also been involved in several successful research proposals.
GRANTS AND PROJECTS
1. Data Science and Analytics Training and Engagement Services for Business, (08/2016 – 07/2017, budget £50k, Higher Education Innovation Fund, Co-I).
2. Grant for Grant in predictive analysis of complex networks – building a Network of Excellence and Programme of Work (01/03/2016 – 31/07/2016, budget £2,000, BU, PI); internal grant at Bournemouth University.
3. ENGINE: European research centre of Network intelliGence for INnovation Enhancement (06/2013 – 05/2017, budget 4.731 Mio. EUR, European Commision, lead at partner organisation). The goal of this project, coordinated by the Wroclaw University of Technology, is to create European research centre of Network intelliGence for INnovation Enhancement.
4. iCANS initiative – interdisciplinary Complex Adaptive Networks and Systems Theory and Applications (05/2012 – 07/2012, budget £2,000, KCL, PI) internal grant at King’s College London, part of EPSRC Bridging the Gaps Interdisciplinary Informatics grant, project leader.
The main goal of this initiative is to enhance cross-disciplinary research at KCL by recognising the links between different research groups and individual researchers in the area of complex networked systems as well as organising meetings and discussion panels to develop new ideas for joint research.
5. The Computational Intelligence Platform for Evolving and Robust Predictive Systems (INFER) project (07/2010–06/2014; budget 1.55 Mio. EUR, European Commission, BU Transfer of Knowledge Coordinator). Project was funded by the European Commission within the Marie Curie Industry and Academia Partnerships & Pathways (IAPP) programme.
Project partners are Evonik Industries from Germany, Research and Engineering Centre (REC) from Poland, and the Smart Technology Research Centre of Bournemouth University in the UK. The project focuses on pervasively adaptive software systems for the development of a modular computational INtelligence software platform For Evolving and Robust predictive systems applicable in various commercial settings and industries. The main innovation of the project is a novel type of environment in which the “fittest” predictive model for whatever purpose will emerge either autonomously or by user high-level goal-related assistance and feedback. I acted as the Transfer of Knowledge coordinator at Bournemouth University side and I was also involved in two research tasks: “Advanced software engineering” and “Complexity research”.
6. GRASP# – Groups, Relationships and Activities of Suspected Persons; Analiza otoczenia spolecznego oraz powi?za? sieciowych osób poszukiwanych i podejrzanych o popelnienie przest?pstwa (10/2009–10/2011, budget £283,000, Polish Ministry of Science and Higher Education, Co-I)
The research and developmental grant from the Polish Ministry of Science and Higher Education. I was a project co-investigator and a member of the project steering committee. The aim of the project was to investigate and analyse the social connections and characteristics of people accused and suspected of committing crime.
7. IT SOA – Service-Oriented Architectures; Nowe technologie informacyjne dla elektronicznej gospodarki i spolecze?stwa informacyjnego oparte na paradygmacie SOA (01/2009–12/2012, budget £8.480 Mio., Polish Ministry of Science and Higher Education, researcher)
The project was developed within the Regional Operational Programme Innovative Economy 2008-2013. I was a member of the project team that was responsible for development of the model and methods for decision support system in the service oriented knowledge utility system.
8. An individual research grant from the Polish Ministry of Science and Higher Education (09/2008–11/2009, budget £8,000, Polish Ministry of Science and Higher Education, PI)
The title of the grant was “A method for analysis of node position in the network of internet users”, number N516 264935. I was a principal investigator of this project.
9. SNAP – Social Network Analysis Platform (02/2008–01/2009, budget £2,000, WRUT, PI)
The grant obtained from the vice-chancellor of the Wroclaw University of Technology. I was the project manager of the “SNAP – Social Network Analysis Platform” project, which was developed by the members of the DaniE group and the purpose of which was to facilitate research on different large social networks.
10. Grant for grants (04/2008–09/2008, budget £15,000, Polish Ministry of Science and Higher Education, Co-I)
The grant that aimed at providing funds for the proposal preparation of “Advanced Methods in Collaborative Knowledge Acquisition and Processing” project proposal within the EU FP7 People programme.
11. Nature-inspired Smart Information Systems – Coordination Action project within EU FP6 (11/2005–01/2008, budget 1 Mio. EUR, European Commission, researcher)
I acted as a member of the Nature-inspired Data Technology (NiDT) focus group within this Network of Excellence. I attended the meetings of the network of excellence at Majorca (06/2006), Tenerife (11/2006), and Malta (11/2007).
Can supervise: YES
My main areas of research are complex networked systems, and analysis of their dynamics and evolution, as well as predictive, adaptive modelling of networked systems. I also recently started research in a new direction – the application of machine learning approaches to networked, dynamical systems. Perfect example of such systems is social network, a concept that we all know very well as each of us is a part of one global network. This network is created by people and the interactions between them. We constantly create connections both in the real world (at home, school, office) and in the rapidly growing online world (Facebook, YouTube, Twitter, Flickr). In my research I investigate those systems, their characteristics and how they change over time. Examples of very interesting questions worth investigating are e.g. what causes that when we work together we can achieve more than when we work individually (concepts known as collective intelligence and emerging behaviour) or what makes that some of the videos, pictures, stories spreads through social network so quickly (known as viral chains).
a) Intelligent analysis of large complex networks – the networks that are in the area of my interest are extracted from large datasets obtained from telecommunication companies (British Telecom plc – BT), e-mail servers (WUT, Enron), multimedia sharing systems (Flickr), etc. The first research on investigating and analysing social networks was conducted as part of the EU FP6 Coordination Action project on Nature-inspired Smart Information Systems (11/2005 – 01/2008) where I acted as a member of the Nature-inspired Data Technology focus group. I presented the research results at the NiSIS symposia at Majorca (06/2006), Tenerife (11/2006), and Malta (11/2007).
b) Dynamics and predictive modelling of complex networked systems – this is area that currently becomes the main field of my research efforts. The conducted research is concerned with discovering patterns in nodes’ behaviours and the interactions between them. The analysis of these patterns and their changes in time allows prediction of the future behaviour of nodes and their relations. One of the ways to model the network dynamics is the application of methods based on the molecular modelling concept and other physically-inspired methods. Another approach that I investigate is the application of machine learning methods to infer and predict the future structure and characteristics of network.
c) Network motifs method in social networks – Network motifs are small subgraphs that reflect local network topology and were shown to be useful for creating profiles that reveal several properties of the network. The outcomes of my research have revealed that motif analysis enables the effective investigation of both network structure and patterns of interactions between nodes within the network. In addition, the analysis of network motifs dynamics can be utilized in detecting and exploration of changes in complex network structures.
d) Multirelational social networks – these are the networks in which more than one type of relationship exists. Different types of relationships can emerge from various communication channels, i.e. based on each communication channel separate relation that can be also called a layer of a network is created. The relationships are extracted from the users activities and if in the system the knowledge about more than one kind of activity is gathered then more than one type of connection can be defined. Different layers can be also built upon various nature of the connections between users, e.g. co-workers, family members, friends. The systems that can be used in such analysis are the multimedia sharing systems such as Flickr or YouTube, which are typical examples of Web 2.0 systems. In my research I have investigated such systems as Flickr, Vimeo, ExtraDom, and recently Badoo.
e) Evaluation of a user position in a social network – during my PhD I conduct research on assessing a position of individuals in networked systems. The position is calculated based on the users activities and their interactions. The method and appropriate algorithms were developed and number of experiments on real-world networks was carried out. The research on evaluating user position was supported by the individual grant that I obtained from the Polish Ministry of Science and Higher Education (06/2008 – 12/2009).
f) Modelling of complex adaptive software systems – the main challenge of the current software systems is to build the systems that will be able to adapt to the changing external environment. The research conducted within the Advanced Software Engineering task within INFER project was focused on developing architectures for complex adaptive systems. I worked on that task when I was employed at BU.
I have taught different units at Undergraduate and Master levels and I have supervised several Undergraduate and Master Projects.
1. Bournemouth University, United Kingdom (2015-2017)
- Postgraduate Framework
2016/2017 – Advanced Data Management, 20 students, unit leader
- Undergraduate Framework
2015/2016 – Data Management, laboratories, 100 students
2015/2016 – Project Management and Teamwork, 180 students, unit leader
2. King’s College London, United Kingdom (2011-2015)
- Postgraduate Framework
2012/2013, 2013/2014 – Project Management, 105 students, module leader
- Undergraduate Framework
2011/2012, 2012/2013 – Data Structures, 180 students, module co-leader
3. Bournemouth University, United Kingdom (2010-2011)
- Postgraduate Framework (Master Course in Information Technology)
2010/2011 – Software Project Management, 13 students, unit leader
2010/2011 – Business System Design, 13 students, unit leader
- Undergraduate Framework
2010/2011 – Data Management, laboratories, 160 students
4. Wroclaw University of Technology, Poland (2006-2009)
- Master Framework (Master Course in Computer Science)
2008/2009 and 2009/2010 – Digital Image Processing, laboratories, lab leader, 45 students
2007/2008 – Data warehouses and data mining, laboratories, lab leader, 30 students
2007/2008 – Interactive web-based multimedia information systems design, laboratories, lab leader, 15 students
2007/2008 – Intelligent information systems, seminars, lab leader, 45 students
- Undergraduate Framework
2009/2010 – Databases, seminars, lab leader, 60 students
2007/2008 and 2008/2009 – Basics of Coding and cryptography, seminars, lab leader, 120 students
Zhang, J, McBurney, P & Musial, K 2018, 'Convergence of trading strategies in continuous double auction markets with boundedly-rational networked traders', Review of Quantitative Finance and Accounting, vol. 50, no. 1, pp. 301-352.View/Download from: UTS OPUS or Publisher's site
© 2017, Springer Science+Business Media New York. This paper considers the convergence of trading strategies among artificial traders connected to one another in a social network and trading in a continuous double auction financial marketplace. Convergence is studied by means of an agent-based simulation model called the Social Network Artificial stoCk marKet model. Six different canonical network topologies (including no-network) are used to represent the possible connections between artificial traders. Traders learn from the trading experiences of their connected neighbours by means of reinforcement learning. The results show that the proportions of traders using particular trading strategies are eventually stable. Which strategies dominate in these stable states depends to some extent on the particular network topology of trader connections and the types of traders.
Kendrick, L, Musial, K & Gabrys, B 2018, 'Change point detection in social networksCritical review with experiments', Computer Science Review, vol. 29, pp. 1-13.View/Download from: UTS OPUS or Publisher's site
© 2018 Elsevier Inc. Change point detection in social networks is an important element in developing the understanding of dynamic systems. This complex and growing area of research has no clear guidelines on what methods to use or in which circumstances. This paper critically discusses several possible network metrics to be used for a change point detection problem and conducts an experimental, comparative analysis using the Enron and MIT networks. Bayesian change point detection analysis is conducted on different global graph metrics (Size, Density, Average Clustering Coefficient, Average Shortest Path) as well as metrics derived from the Hierarchical and Block models (Entropy, Edge Probability, No. of Communities, Hierarchy Level Membership). The results produced the posterior probability of a change point at weekly time intervals that were analysed against ground truth change points using precision and recall measures. Results suggest that computationally heavy generative models offer only slightly better results compared to some of the global graph metrics. The simplest metrics used in the experiments, i.e. nodes and links numbers, are the recommended choice for detecting overall structural changes.
Qin, M, Lei, K, Gabrys, B & Musial-Gabrys, K 2018, 'Adaptive community detection incorporating topology and content in social networks', Knowledge-Based Systems, vol. 161, pp. 342-356.View/Download from: UTS OPUS or Publisher's site
© 2018 In social network analysis, community detection is a basic step to understand the structure and function of networks. Some conventional community detection methods may have limited performance because they merely focus on the networks' topological structure. Besides topology, content information is another significant aspect of social networks. Although some state-of-the-art methods started to combine these two aspects of information for the sake of the improvement of community partitioning, they often assume that topology and content carry similar information. In fact, for some examples of social networks, the hidden characteristics of content may unexpectedly mismatch with topology. To better cope with such situations, we introduce a novel community detection method under the framework of non-negative matrix factorization (NMF). Our proposed method integrates topology as well as content of networks and has an adaptive parameter (with two variations) to effectively control the contribution of content with respect to the identified mismatch degree. Based on the disjoint community partition result, we also introduce an additional overlapping community discovery algorithm, so that our new method can meet the application requirements of both disjoint and overlapping community detection. The case study using real social networks shows that our new method can simultaneously obtain the community structures and their corresponding semantic description, which is helpful to understand the semantics of communities. Related performance evaluations on both artificial and real networks further indicate that our method outperforms some state-of-the-art methods while exhibiting more robust behavior when the mismatch between topology and content is observed.
De Meo, P, Musial-Gabrys, K, Rosaci, D, Sarne, GML & Aroyo, L 2017, 'Using centrality measures to predict helpfulness-based reputation in trust networks', ACM Transactions on Internet Technology, vol. 17, no. 1.View/Download from: Publisher's site
© 2017 ACM 1533-5399/2017/02-ART8 15.00. In collaborativeWeb-based platforms, user reputation scores are generally computed according to two orthogonal perspectives: (a) helpfulness-based reputation (HBR) scores and (b) centrality-based reputation (CBR) scores. InHBR approaches, the most reputable users are those who post the most helpful reviews according to the opinion of the members of their community. In CBR approaches, a "who-Trusts-whom" network-known as a trust network-is available and the most reputable users occupy the most central position in the trust network, according to some definition of centrality. The identification of users featuring large HBR scores is one of the most important research issue in the field of Social Networks, and it is a critical success factor of many Web-based platforms like e-marketplaces, product review Web sites, and question-And-Answering systems. Unfortunately, user reviews/ratings are often sparse, and this makes the calculation of HBR scores inaccurate. In contrast, CBR scores are relatively easy to calculate provided that the topology of the trust network is known. In this article, we investigate if CBR scores are effective to predict HBR ones, and, to perform our study, we used real-life datasets extracted from CIAO and Epinions (two product review Web sites) andWikipedia and applied five popular centrality measures-Degree Centrality, Closeness Centrality, Betweenness Centrality, PageRank and Eigenvector Centrality-to calculate CBR scores. Our analysis provides a positive answer to our research question: CBR scores allow for predicting HBR ones and Eigenvector Centrality was found to be the most important predictor. Our findings prove that we can leverage trust relationships to spot those users producing the most helpful reviews for the whole community.
Gao, F, Musial, K, Cooper, C & Tsoka, S 2015, 'Link prediction methods and their accuracy for different social networks and network metrics', Scientific Programming, vol. 2015.View/Download from: Publisher's site
Copyright © 2015 Fei Gao et al. Currently, we are experiencing a rapid growth of the number of social-based online systems. The availability of the vast amounts of data gathered in those systems brings new challenges that we face when trying to analyse it. One of the intensively researched topics is the prediction of social connections between users. Although a lot of effort has been made to develop new prediction approaches, the existing methods are not comprehensively analysed. In this paper we investigate the correlation between network metrics and accuracy of different prediction methods.We selected six time-stamped real-world social networks and ten most widely used link prediction methods. The results of the experiments show that the performance of some methods has a strong correlation with certain network metrics. We managed to distinguish "prediction friendly" networks, for which most of the prediction methods give good performance, as well as "prediction unfriendly" networks, for which most of the methods result in high prediction error. Correlation analysis between networkmetrics and prediction accuracy of prediction methodsmay formthe basis of ametalearning system where based on network characteristics it will be able to recommend the right prediction method for a given network.
Musial, K, Bródka, P, Kazienko, P & Gaworecki, J 2014, 'Extraction of multilayered social networks from activity data.', TheScientificWorldJournal, vol. 2014, p. 359868.View/Download from: Publisher's site
The data gathered in all kinds of web-based systems, which enable users to interact with each other, provides an opportunity to extract social networks that consist of people and relationships between them. The emerging structures are very complex due to the number and type of discovered connections. In web-based systems, the characteristic element of each interaction between users is that there is always an object that serves as a communication medium. This can be, for example, an e-mail sent from one user to another or post at the forum authored by one user and commented on by others. Based on these objects and activities that users perform towards them, different kinds of relationships can be identified and extracted. Additional challenge arises from the fact that hierarchies can exist between objects; for example, a forum consists of one or more groups of topics, and each of them contains topics that finally include posts. In this paper, we propose a new method for creation of multilayered social network based on the data about users activities towards different types of objects between which the hierarchy exists. Due to the flattening, preprocessing procedure of new layers and new relationships in the multilayered social network can be identified and analysed.
Budka, M, Juszczyszyn, K, Musial, K & Musial, A 2013, 'Molecular model of dynamic social network based on e-mail communication', Social Network Analysis and Mining, vol. 3, no. 3, pp. 543-563.View/Download from: Publisher's site
© 2013, The Author(s). In this work we consider an application of physically inspired sociodynamical model to the modelling of the evolution of email-based social network. Contrary to the standard approach of sociodynamics, which assumes expressing of system dynamics with heuristically defined simple rules, we postulate the inference of these rules from the real data and their application within a dynamic molecular model. We present how to embed the n-dimensional social space in Euclidean one. Then, inspired by the Lennard-Jones potential, we define a data-driven social potential function and apply the resultant force to a real e-mail communication network in a course of a molecular simulation, with network nodes taking on the role of interacting particles. We discuss all steps of the modelling process, from data preparation, through embedding and the molecular simulation itself, to transformation from the embedding space back to a graph structure. The conclusions, drawn from examining the resultant networks in stable, minimum-energy states, emphasize the role of the embedding process projecting the non–metric social graph into the Euclidean space, the significance of the unavoidable loss of information connected with this procedure and the resultant preservation of global rather than local properties of the initial network. We also argue applicability of our method to some classes of problems, while also signalling the areas which require further research in order to expand this applicability domain.
Musial, K, Budka, M & Blysz, W 2013, 'Understanding the Other Side - The Inside Story of the INFER Project', Smart Innovation, Systems and Technologies, vol. 18, pp. 1-9.View/Download from: Publisher's site
In the last few years, the collaboration between research institutions and industry has become a well established process. Transfer of Knowledge (ToK) is required to accelerate the development of both sides and to enable them to unlock their full potential. European Commission within the Marie Curie Industry and Academia Partnerships & Pathways (IAPP) programme supports the cooperation between these two sectors at the international scale by funding research projects that as one of the objectives aim at enhancing human mobility. IAPP projects offer people from different institutions the possibility to move sector and country in order to provide, absorb and implement new knowledge in a professional industrial-academic environment. In this paper, one of such projects is presented and both academia and industry perspectives in regard to opportunities and challenges in Transfer of Knowledge are described. Computational Intelligence Platform for Evolving and Robust Predictive Systems (INFER) is the IAPP project that serves as a case study for this paper. © Springer-Verlag Berlin Heidelberg 2013.
Musial, K, Budka, M & Juszczyszyn, K 2013, 'Creation and growth of online social network: How do social networks evolve?', World Wide Web, vol. 16, no. 4, pp. 421-447.View/Download from: Publisher's site
Social networks are an example of complex systems consisting of nodes that can interact with each other and based on these activities the social relations are defined. The dynamics and evolution of social networks are very interesting but at the same time very challenging areas of research. In this paper the formation and growth of one of such structures extracted from data about human activities within online social networking system is investigated. Dynamics of both local and global characteristics are studied. Analysis of the dynamics of the network growth showed that it changes over time-from random process to power-law growth. The phase transition between those two is clearly visible. In general, node degree distribution can be described as the scale-free but it does not emerge straight from the beginning. Social networks are known to feature high clustering coefficient and friend-of-a-friend phenomenon. This research has revealed that in online social network, although the clustering coefficient grows over time, it is lower than expected. Also the friend-of-a-friend phenomenon is missing. On the other hand, the length of the shortest paths is small starting from the beginning of the network existence so the small-world phenomenon is present. The unique element of the presented study is that the data, from which the online social network was extracted, represents interactions between users from the beginning of the social networking site existence. The system, from which the data was obtained, enables users to interact using different communication channels and it gives additional opportunity to investigate multi-relational character of human relations. © 2012 The Author(s).
Musiał, K & Kazienko, P 2013, 'Social networks on the Internet', World Wide Web, vol. 16, no. 1, pp. 31-72.View/Download from: Publisher's site
The rapid development and expansion of the Internet and the social-based services comprised by the common Web 2.0 idea provokes the creation of the new area of research interests, i. e. social networks on the Internet called also virtual or online communities. Social networks can be either maintained and presented by social networking sites like MySpace, LinkedIn or indirectly extracted from the data about user interaction, activities or achievements such as emails, chats, blogs, homepages connected by hyperlinks, commented photos in multimedia sharing system, etc. A social network is the set of human beings or rather their digital representations that refer to the registered users who are linked by relationships extracted from the data about their activities, common communication or direct links gathered in the internet-based systems. Both digital representations named in the paper internet identities as well as their relationships can be characterized in many different ways. Such diversity yields for building a comprehensive and coherent view onto the concept of internet-based social networks. This survey provides in-depth analysis and classification of social networks existing on the Internet together with studies on selected examples of different virtual communities. © 2012 The Author(s).
Bródka, P, Kazienko, P, Musiał, K & Skibicki, K 2012, 'Analysis of Neighbourhoods in Multi-layered Dynamic Social Networks', International Journal of Computational Intelligence Systems, vol. 5, no. 3, pp. 582-596.View/Download from: Publisher's site
Social networks existing among employees, customers or other types of users of various IT systems have become one of the research areas of growing importance. Data about people and their interactions that exist in social media, provides information about many different types of relationships within one network. Analysing this data one can obtain knowledge not only about the structure and characteristics of the network but it also enables to understand the semantic of human relations.Each social network consists of nodes - social entities and edges linking pairs of nodes. In regular, one-layered networks, two nodes - i.e. people are connected with a single edge whereas in the multi-layered social networks, there may be many links of different types for a pair of nodes. Most of the methods used for social network analysis (SNA) may be applied only to one-layered networks. Thus, some new structural measures for multi-layered social networks are proposed in the paper. This study focuses on definitions and analysis of cross-layer clustering coefficient, cross-layer degree centrality and various versions of multi-layered degree centralities. Authors also investigated the dynamics of multi-layered neighbourhood. The evaluation of the presented concepts on the real-world dataset is presented. The measures proposed in the paper may directly be used to various methods for collective classification, in which nodes are assigned to labels according to their structural input features. © 2012 Copyright the authors.
Kazienko, P, Musiał, K & Kajdanowicz, T 2011, 'Multidimensional social network in the social recommender system', IEEE Transactions on Systems, Man, and Cybernetics Part A:Systems and Humans, vol. 41, no. 4, pp. 746-759.View/Download from: Publisher's site
All online sharing systems gather data that reflects users' collective behavior and their shared activities. This data can be used to extract different kinds of relationships which can be grouped into layers and which are basic components of the multidimensional social network (MSN) proposed in the paper. The layers are created on the basis of two types of relations between humans, i.e., direct and object-based ones which, respectively, correspond to either social or semantic links between individuals. For better understanding of the complexity of the social network structure, layers and their profiles were identified and studied on two, spanned in time, snapshots of the 'Flickr' population. Additionally, for each layer, a separate strength measure was proposed. The experiments on the 'Flickr' photo sharing system revealed that the relationships between users result either from semantic links between objects they operate on or from social connections of these users. Moreover, the density of the social network increases in time. The second part of this paper is devoted to building a social recommender system that supports the creation of new relations between users in a multimedia sharing system. Its main goal is to generate personalized suggestions that are continuously adapted to users' needs depending on the personal weights assigned to each layer in the MSN. The conducted experiments confirmed the usefulness of the proposed model. © 2011 IEEE.
Budka, M, Gabrys, B & Musial, K 2011, 'On accuracy of PDF divergence estimators and their applicability to representative data sampling', Entropy, vol. 13, no. 7, pp. 1229-1266.View/Download from: Publisher's site
Generalisation error estimation is an important issue in machine learning. Cross-validation traditionally used for this purpose requires building multiple models and repeating the whole procedure many times in order to produce reliable error estimates. It is however possible to accurately estimate the error using only a single model, if the training and test data are chosen appropriately. This paper investigates the possibility of using various probability density function divergence measures for the purpose of representative data sampling. As it turned out, the first difficulty one needs to deal with is estimation of the divergence itself. In contrast to other publications on this subject, the experimental results provided in this study show that in many cases it is not possible unless samples consisting of thousands of instances are used. Exhaustive experiments on the divergence guided representative data sampling have been performed using 26 publicly available benchmark datasets and 70 PDF divergence estimators, and their results have been analysed and discussed. © 2011 by the authors; licensee MDPI, Basel, Switzerland.
Juszczyszyn, K, Kazienko, P & Musiał, K 2010, 'Personalized ontology-based recommender systems for multimedia objects', Studies in Computational Intelligence, vol. 289, pp. 275-292.View/Download from: Publisher's site
A framework for recommendation of multimedia objects based on processing of individual ontologies is proposed in the chapter. The recommendation process takes into account similarities calculated both between objects' and users' ontologies, which reflect the social and semantic features existing in the system. The ontologies, which are close to the current context, provide a list of suggestions presented to the user. Each user in the system possesses its own Personal Agent that performs all necessary online tasks. Personal Agents co-operate each other and enrich lists of possible recommendations. The system was developed for the use inthe Flickr multimedia sharing system. © 2010 Springer-Verlag Berlin Heidelberg.
Juszczyszyn, K, Musiał, K, Kazienko, P & Gabrys, B 2009, 'Temporal changes in local topology of an email-based social network', Computing and Informatics, vol. 28, no. 6, pp. 763-779.
The dynamics of complex social networks has become one of the research areas of growing importance. The knowledge about temporal changes of the network topology and characteristics is crucial in networked communication systems in which accurate predictions are important. The local network topology can be described by the means of network motifs which are small subgraphs - usually containing from 3 to 7 nodes. They were shown to be useful for creating profiles that reveal several properties of the network. In this paper, the time-varying characteristics of social networks, such as the number of nodes and edges as well as clustering coefficients and different centrality measures are investigated. At the same time, the analysis of three-node motifs (triads) was used to track the temporal changes in the structure of a large social network derived from e-mail communication between university employees. We have shown that temporal changes in local connection patterns of the social network are indeed correlated with the changes in the clustering coefficient as well as various centrality measures values and are detectable by means of motifs analysis. Together with robust sampling network motifs can provide an appealing way to monitor and assess temporal changes in large social networks.
Musiał, K, Juszczyszyn, K & Kazienko, P 2008, 'Ontology-based recommendation in multimedia sharing systems', Systems Science, vol. 34, no. 1, pp. 97-106.
In this paper, a new framework for recommendation of multimedia objects in the environment of the multimedia sharing system has been proposed. It uses two kinds of individual ontologies, one is created for multimedia objects and the second one for system users. The final recommendation process takes into account similarities calculated both between objects' and users' ontologies. These individual ontologies respect all the social and semantic features existing in the system. The entire recommender framework was developed for the use in Flickr, a typical photo sharing system.
Kazienko, P & Musial, K 2007, 'On utilising social networks to discover representatives of human communities', International Journal of Intelligent Information and Database Systems, vol. 1, no. 3-4, pp. 293-310.
: Virtual human communities that exist on the internet reflect social relationships between people. There is a great need to find important individuals, a set of people who would represent larger communities. These people would be able to perform specific tasks or could become a target group for marketing or advertising purposes. The new research on representative discovery for human communities is presented in this paper. Its main goal is to improve the process of target group selection by adding the social elements derived from the behaviours of people. The entire selection process considered in the paper is called human filtering. © 2007 Inderscience Enterprises Ltd.
Kazienko, P & Musiał, K 2006, 'Recommendation framework for online social networks', Studies in Computational Intelligence, vol. 23, pp. 111-120.View/Download from: Publisher's site
The recommendation framework that supports the creation of new interpersonal relationships within the social networks is presented in the paper. It integrates many sources of data in order to generate the relevant personalized recommendations for network members. The unique social filtering techniques and measures of the activity and strength of relationship are encompassed by the framework. © Springer-Verlag Berlin Heidelberg 2006.
Jin, D, Liu, Z, He, D, Gabrys, B & Musial, K 2018, 'Robust detection of communities with multi-semantics in large attributed networks', Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), pp. 362-376.View/Download from: UTS OPUS or Publisher's site
© 2018, Springer Nature Switzerland AG. In this paper, we are interested in how to explore and utilize the relationship between network communities and semantic topics in order to find the strong explanatory communities robustly. First, the relationship between communities and topics displays different situations. For example, from the viewpoint of semantic mapping, their relationship can be one-to-one, one-to-many or many-to-one. But from the standpoint of underlying community structures, the relationship can be consistent, partially consistent or completely inconsistent. Second, it will be helpful to not only find communities more precise but also reveal the communities' semantics that shows the relationship between communities and topics. To better describe this relationship, we introduce the transition probability which is an important concept in Markov chain into a well-designed nonnegative matrix factorization framework. This new transition probability matrix with a suitable prior which plays the role of depicting the relationship between communities and topics can perform well in this task. To illustrate the effectiveness of the proposed new approach, we conduct some experiments on both synthetic and real networks. The results show that our new method is superior to baselines in accuracy. We finally conduct a case study analysis to validate the new method's strong interpretability to detected communities.
© 2018 The Author(s). Networks are everywhere and their many types, including social networks, the Internet, food webs etc., have been studied for the last few decadeS. However, in real-world networks, it's hard to find examples that can be easily comparable, i.e. have the same density or even number of nodes and edgeS. We propose a flexible and extensible Netsim framework to understand how properties in different types of networks change with varying number of edges and verticeS. Our approach enables to simulate three classical network models (random, small-world and scale-free) with easily adjustable model parameters and network size. To be able to compare different networks, for a single experimental setup we kept the number of edges and vertices fixed across the modelS. To understand how they change depending on the number of nodes and edges we ran over 30,000 simulations and analysed different network characteristics that cannot be derived analytically. Two of the main findings from the analysis are that the average shortest path does not change with the density of the scale-free network but changes for small-world and random networks; the apparent difference in mean betweenness centrality of the scale-free network compared with random and small-world networkS.
Qin, M, Jin, D, He, D, Gabrys, B & Musial, K 2017, 'Adaptive community detection incorporating topology and content in social networks', Proceedings of the 2017 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining, ASONAM 2017, pp. 675-682.View/Download from: Publisher's site
© 2017 Association for Computing Machinery. In social network analysis, community detection is a basic step to understand the structure, function and semantics of networks. Some conventional community detection methods may have limited performance because they merely focus on topological structure of networks. In addition to topology, content information is another significant aspect of social networks. Some state-of-the-art methods started to combine these two aspects of information, but they often assume that topology and content share the same characteristics. However, for some examples of social networks, content may mismatch with topological structure. In order to better cope with such situations, we introduce a novel community detection method under the framework of nonnegative matrix factorization (NMF). Our proposed method integrates topology and content of networks, and introduces a novel adaptive parameter for controlling the contribution of content with respect to the identified mismatch degree between the topological and content information. The case study using real social networks show that our new method can simultaneously obtain community partition and the corresponding semantic descriptions. Experiments on both artificial networks and real social networks further indicate that our method outperforms some state-of-the-art methods while exhibiting more robust behaviour when the mismatch topological and content information is observed.
Gao, F, Musial, K & Gabrys, B 2017, 'A community bridge boosting social network link prediction model', Proceedings of the 2017 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining, ASONAM 2017, pp. 683-689.View/Download from: Publisher's site
© 2017 Association for Computing Machinery. Link prediction in social networks is a very challenging research problem. The majority of existing approaches are based on the assumption that a given network evolves following a single phenomenon, e.g.'rich get richer' or'friend of my friend is my friend'. However, dynamics of network dynamic changes over time and different parts of the network evolve in different manner. Because of that, we hypothesise that the prediction accuracy can be improved by providing different treatment to different nodes and links. Building on that assumption, we propose a Community Bridge Boosting Prediction Model (CBBPM) that treats certain bridge nodes differently depending on their structural position. For such bridge nodes their similarity score obtained using traditional link-based prediction methods is boosted. By doing so the importance of these nodes is increased and at the same time ensuring that the CBBPM can be used with any existing link prediction method. Our experimental results show that such bridge node similarity boosting mechanism can improve the accuracy of traditional link prediction methods.
Venkata, SK, Musial, K, Mahmoud, S & Keppens, J 2017, '[title field missing]', Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), pp. 157-169.View/Download from: Publisher's site
© Springer International Publishing AG 2017. With the growing number of applications that require large data transfers from distributed databases, there is a great need for efficient distributed data caching methods. It is essential that data is cached at the best and optimal locations between users and data stores. Cache management should consider patterns about data usage and make dynamic decisions to place data across cache units. In this paper, we have modelled the distributed data caching mechanism using multi-agent system allowing to test strategies and algorithms for data placement that later can be incorporated in the real life applications. Subsequently, we demonstrate the application of this system to study various distributed coordination strategies for identifying effective data placement and thus improving overall cache performance. This study is significant for distributed system applications.
Venkata, SK, Musial, K, Mahmoud, S & Keppens, J 2017, 'Demonstration: Multi-agent system for distributed cache maintenance', Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), pp. 364-368.View/Download from: Publisher's site
© Springer International Publishing AG 2017. Innovations in science and technology is increasing the demand on huge data transfers and hence number of data caches. In this paper, we consider the community caching solution, CommCache, where many groups of users are working together on related projects distributed all over the world. We demonstrate the use of proactive caches for data placement problem with the help of multi-agent coordination.
Maggs, CA & Musial-Gabrys, K 2017, 'LINNEAN SYSTEMATICS IN THE AGE OF BIG DATA', PHYCOLOGIA, INT PHYCOLOGICAL SOC, pp. 123-124.
Kajdanowicz, T, Michalski, R, Musial, K & Kazienko, P 2016, 'Learning in unlabeled networks - An active learning and inference approach', AI Communications, pp. 123-148.View/Download from: Publisher's site
© 2016 - IOS Press and the authors. All rights reserved. The task of determining labels of all network nodes based on the knowledge about network structure and labels of some training subset of nodes is called the within-network classification. It may happen that none of the labels of the nodes is known and additionally there is no information about number of classes (types of labels) to which nodes can be assigned. In such a case a subset of nodes has to be selected for initial label acquisition. The question that arises is: labels of which nodes should be collected and used for learning in order to provide the best classification accuracy for the whole network?. Active learning and inference is a practical framework to study this problem. In this paper, set of methods for active learning and inference for within-network classification is proposed and validated. The utility score calculation for each node based on network structure is the first step in the entire process. The scores enable to rank the nodes. Based on the created ranking, a set of nodes, for which the labels are acquired, is selected (e.g. by taking top or bottom N from the ranking). The new measure-neighbour methods proposed in the paper suggest not obtaining labels of nodes from the ranking but rather acquiring labels of their neighbours. The paper examines 29 distinct formulations of utility score and selection methods reporting their impact on the results of two collective classification algorithms: Iterative Classification Algorithm (ICA) and Loopy Belief Prorogation (LBP). We advocate that the accuracy of presented methods depends on the structural properties of the examined network. We claim that measure-neighbour methods will work better than the regular methods for networks with higher clustering coefficient and worse than regular methods for networks with low clustering coefficient. According to our hypothesis, based on clustering coefficient of a network we are able to recommend approp...
Abdullaev, S, McBurney, P & Musial, K 2016, 'Pricing options with portfolio-holding trading agents in direct double auction', Frontiers in Artificial Intelligence and Applications, pp. 1754-1755.View/Download from: Publisher's site
© 2016 The Authors and IOS Press. Options constitute integral part of modern financial trades, and are priced according to the risk associated with buying or selling certain asset in future. Financial literature mostly concentrates on risk-neutral methods of pricing options such as Black-Scholes model. However, it is an emerging field in option pricing theory to use trading agents with utility functions to determine the option's potential payoff for the agent. In this paper, we use one of such methodologies developed by Othman and Sandholm to design portfolio-holding agents that are endowed with popular option portfolios such as bullish spread, butterfly spread, straddle, etc to price options. Agents use their portfolios to evaluate how buying or selling certain option would change their current payoff structure, and form their orders based on this information. We also simulate these agents in a multi-unit direct double auction. The emerging prices are compared to risk-neutral prices under different market conditions. Through an appropriate endowment of option portfolios to agents, we can also mimic market conditions where the population of agents are bearish, bullish, neutral or non-neutral in their beliefs.
Gao, F & Musial-Gabrys, K 2016, 'Hybrid structure-based link prediction model', Proceedings of the 2016 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining, ASONAM 2016, pp. 1221-1228.View/Download from: Publisher's site
© 2016 IEEE. In network science several topology-based link prediction methods have been developed so far. The classic social network link prediction approach takes as an input a snapshot of a whole network. However, with human activities behind it, this social network keeps changing. In this paper, we consider link prediction problem as a time-series problem and propose a hybrid link prediction model that combines eight structure-based prediction methods and self-adapts the weights assigned to each included method. To test the model, we perform experiments on two real world networks with both sliding and growing window scenarios. The results show that our model outperforms other structure-based methods when both precision and recall of the prediction results are considered.
Venkata, SK, Keppens, J & Musial, K 2016, 'Agent based simulation to evaluate adaptive caching in distributed databases', Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), pp. 455-462.View/Download from: Publisher's site
© Springer International Publishing Switzerland 2016. Caching frequently used data is a common practice to improve query performance in database systems. But traditional algorithms used for cache management prove to be insufficient in distributed environment where groups of users require similar or related data from multiple databases. Repeated data transfers can become a bottleneck leading to long query response time and high resource utilization. Our work focuses on adaptive algorithms to decide on optimal grain of data to be cached and cache refreshment techniques to reduce data transfers. In this paper, we present agent based simulation to investigate and in consequence improve cache management in the distributed database environment. Dynamic grain size and decisions on cache refreshment are made as a result of coordination and interaction between agents. Initial results show better response time and higher data availability compared to traditional caching techniques.
Abdullaev, S, Mcburney, P & Musial, K 2015, 'Direct exchange mechanisms for option pricing', Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), pp. 269-284.View/Download from: Publisher's site
© Springer International Publishing Switzerland 2015. This paper presents the design and simulation of direct exchange mechanisms for pricing European options. It extends McAfee's single-unit double auction to multi-unit format, and then applies it for pricing options through aggregating agent predictions of future asset prices. We will also propose the design of a combinatorial exchange for the simulation of agents using option trading strategies. We present several option trading strategies that are commonly used in real option markets to minimise the risk of future loss, and assume that agents can submit them as a combinatorial bid to the market maker. We provide simulation results for proposed mechanisms, and compare them with existing Black-Scholes model mostly used for option pricing. The simulation also tests the effect of supply and demand changes on option prices. It also takes into account agents with different implied volatility. We also observe how option prices are affected by the agents' choices of option trading strategies.
Bródka, P, Magnani, M & Musial, K 2014, 'Message from SNAA 2014 program chairs', ASONAM 2014 - Proceedings of the 2014 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining, p. xxxiv.View/Download from: Publisher's site
Król, D, Budka, M & Musial, K 2014, 'Simulating the information diffusion process in complex networks using push and pull strategies', Proceedings - 2014 European Network Intelligence Conference, ENIC 2014, pp. 1-8.View/Download from: Publisher's site
© 2014 IEEE. Simulation of information diffusion has recently attracted considerable interest from both academia and industry. One of the main reasons for this is the continuously expanding online world where sharing information is a 'one click' activity. This poses new research challenges related to explaining how the information spreads within and across a variety of online communities. The main goal of this study is to investigate how the information diffusion process can be simulated. The contribution of this paper is two - fold. First, we propose a new generic diffusion - based algorithm that is applicable in a wide range of scenarios. The algorithm is based on a social propagation mechanism, which embraces two most widely employed models: the independent cascade model and the linear threshold model. The second contribution of this research is a systematic empirical assessment of how proposed diffusion simulation strategies are associated with scope and speed of spread over a range of network structures and diffusion parameters.
Kajdanowicz, T, Michalski, R, Musial, K & Kazienko, P 2013, 'Active learning and inference method for within network classification', Proceedings of the 2013 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining, ASONAM 2013, pp. 1299-1306.View/Download from: Publisher's site
In relational learning tasks such as within network classification the main problem arises from the inference of nodes' labels based on the the ground true labels of remaining nodes. The problem becomes even harder if the nodes from initial network do not have any labels assigned and they have to be acquired. However, labels of which nodes should be obtained in order to provide fair classification results? Active learning and inference is a practical framework to study this problem. The method for active learning and inference in within network classification based on node selection is proposed in the paper. Based on the structure of the network it is calculated the utility score for each node, the ranking is formulated and for selected nodes the labels are acquired. The paper examines several distinct proposals for utility scores and selection methods reporting their impact on collective classification results performed on various real-world networks. Copyright 2013 ACM.
Musial, K, Gabrys, B & Buczko, M 2013, 'What kind of network are you? - Using local and global characteristics in network categorisation tasks', Proceedings of the 2013 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining, ASONAM 2013, pp. 1366-1373.View/Download from: Publisher's site
The amount of research done in the area of real- world networked systems is rapidly growing. Everybody knows what six degrees of separation or small-world phenomenon are. Scientists very easily give labels to the networks they analyse. If it has power law node degree distribution then it has to be scale-free network or if there is high clustering coefficient then it must be small-world network. These simplifications, although convenient, are not always very useful from the perspective of understanding phenomena existing within the network. In this paper we decided to go back to the basics and investigate whether analysis of one single measure is enough to describe a network. We analyse both local and global characteristics in order to discover the "true" nature of a network. Not only using local and/or global measures can lead to different classification of a network but we also show how significantly different interpretation can result from analysing the same data by building network models as directed/undirected and/or weighted/binary graphs. Copyright 2013 ACM.
Juszczyszyn, K, Gonczarek, A, Tomczak, JM, Musial, K & Budka, M 2012, 'A probabilistic approach to structural change prediction in evolving social networks', Proceedings of the 2012 IEEE/ACM International Conference on Advances in Social Networks Analysis and Mining, ASONAM 2012, pp. 996-1001.View/Download from: Publisher's site
We propose a predictive model of structural changes in elementary subgraphs of social network based on Mixture of Markov Chains. The model is trained and verified on a dataset from a large corporate social network analyzed in short, one day-long time windows, and reveals distinctive patterns of evolution of connections on the level of local network topology. We argue that the network investigated in such short timescales is highly dynamic and therefore immune to classic methods of link prediction and structural analysis, and show that in the case of complex networks, the dynamic subgraph mining may lead to better prediction accuracy. The experiments were carried out on the logs from the Wroclaw University of Technology mail server. © 2012 IEEE.
Budka, M, Musial, K & Juszczyszyn, K 2012, 'Predicting the evolution of social networks: Optimal time window size for increased accuracy', Proceedings - 2012 ASE/IEEE International Conference on Privacy, Security, Risk and Trust and 2012 ASE/IEEE International Conference on Social Computing, SocialCom/PASSAT 2012, pp. 21-30.View/Download from: Publisher's site
This study investigates the data preparation process for predictive modelling of the evolution of complex networked systems, using an e - mail based social network as an example. In particular, we focus on the selection of optimal time window size for building a time series of network snapshots, which forms the input of chosen predictive models. We formulate this issue as a constrained multi - objective optimization problem, where the constraints are specific to a particular application and predictive algorithm used. The optimization process is guided by the proposed Windows Incoherence Measures, defined as averaged Jensen-Shannon divergences between distributions of a range of network characteristics for the individual time windows and the network covering the whole considered period of time. The experiments demonstrate that the informed choice of window size according to the proposed approach allows to boost the prediction accuracy of all examined prediction algorithms, and can also be used for optimally defining the prediction problems if some flexibility in their definition is allowed. © 2012 IEEE.
Musial, K & Sastry, N 2012, 'Social media - Are they underpinned by social or interest-based interactions?', ACM International Conference Proceeding Series, pp. 1-6.View/Download from: Publisher's site
On many social media and user-generated content sites, users can not only upload content but also create links with other users to follow their activities. It is interesting to ask whether the resulting user-user Followers' Network is based more on social ties, or shared interests in similar content. This paper reports our preliminary progress in answering this question using around five years of data from social video-sharing site vimeo. Many links in the Followers' Network are between users who do not have any videos in common, which would imply the network is not interest-based, but rather has a social character. However, the Followers' Network also exhibits properties unlike other social networks, for instance, clustering co-efficient is low, links are frequently not reciprocated, and users form links across vast geographical distances. In addition, analysis of the relationship strength, calculated as the number of commonly liked videos, people who follow each other and share some "likes" have more video likes in common than the general population. We conclude by speculating on the reasons for these differences and proposals for further work. © 2012 ACM.
Kazienko, P, Musial, K, Kukla, E, Kajdanowicz, T & Bródka, P 2011, 'Multidimensional social network: Model and analysis', Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), pp. 378-387.View/Download from: Publisher's site
A social network is an abstract concept consisting of set of people and relationships linking pairs of humans. A new multidimensional model, which covers three main dimensions: relation layer, time window and group, is proposed in the paper. These dimensions have a common set of nodes, typically, corresponding to human beings. Relation layers, in turn, reflect various relationship types extracted from different user activities gathered in computer systems. The time dimension corresponds to temporal variability of the social network. Social groups are extracted by means of clustering methods and group people who are close to each other. An atomic component of the multidimensional social network is a view - small social sub-network, which is in the intersection of all dimensions. A view describes the state of one social group, linked by one type of relationship (one layer), and derived from one time period. The multidimensional model of a social network is similar to a general concept of data warehouse, in which a fact corresponds to a view. Aggregation possibilities and usage of the model is also discussed in the paper. © 2011 Springer-Verlag Berlin Heidelberg.
Kazienko, P, Kukla, E, Musial, K, Kajdanowicz, T, Bródka, P & Gaworecki, J 2011, 'A generic model for a multidimensional temporal social network', Communications in Computer and Information Science, pp. 1-14.View/Download from: Publisher's site
A comprehensive generic model for a multidimensional, temporal social network is proposed in the paper. It covers three main dimensions: layers, time windows and social groups. All the dimensions share the same set of nodes corresponding to social entities, usually individuals. Layers correspond to different types of relationships between humans, e.g. social and semantic, that can be derived from different human activities in IT systems; time windows reflect the temporal profile of the social network, whereas groups (social communities) are sets of similar humans. The intersection of all dimensions is called a view; it represents the statement of a single social cluster (group) with connections of only one type (from a single layer) and with the snapshot for a given period. Views can be aggregated by one, two or even all three dimensions simultaneously using filtering of dimension instances. Apart from description of the multidimensional model, its applicability is also considered in the paper. © 2011 Springer-Verlag.
Bródka, P, Skibicki, K, Kazienko, P & Musiał, K 2011, 'A degree centrality in multi-layered social network', Proceedings of the 2011 International Conference on Computational Aspects of Social Networks, CASoN'11, pp. 237-242.View/Download from: Publisher's site
Multi-layered social networks reflect complex relationships existing in modern interconnected IT systems. In such a network each pair of nodes may be linked by many edges that correspond to different communication or collaboration user activities. Multi-layered degree centrality for multi-layered social networks is presented in the paper. Experimental studies were carried out on data collected from the real Web 2.0 site. The multi-layered social network extracted from this data consists of ten distinct layers and the network analysis was performed for different degree centralities measures. © 2011 IEEE.
Juszczyszyn, K, Musiał, K & Budka, M 2011, 'Link prediction based on Subgraph evolution in dynamic social networks', Proceedings - 2011 IEEE International Conference on Privacy, Security, Risk and Trust and IEEE International Conference on Social Computing, PASSAT/SocialCom 2011, pp. 27-34.View/Download from: Publisher's site
We propose a new method for characterizing the dynamics of complex networks with its application to the link prediction problem. Our approach is based on the discovery of network subgraphs (in this study: triads of nodes) and measuring their transitions during network evolution. We define the Triad Transition Matrix (TTM) containing the probabilities of transitions between triads found in the network, then we show how it can help to discover and quantify the dynamic patterns of network evolution. We also propose the application of TTM to link prediction with an algorithm (called TTM-predictor) which shows good performance, especially for sparse networks analyzed in short time scales. The future applications and research directions of our approach are also proposed and discussed. © 2011 IEEE.
Juszczyszyn, K, Budka, M & Musiał, K 2011, 'The dynamic structural patterns of social networks based on triad transitions', Proceedings - 2011 International Conference on Advances in Social Networks Analysis and Mining, ASONAM 2011, pp. 581-586.View/Download from: Publisher's site
In modern social networks built from the data collected in various computer systems we observe constant changes corresponding to external events or the evolution of underlying organizations. In this work we present a new approach to the description and quantifying evolutionary patterns of social networks illustrated with the data from the Enron email dataset. We propose the discovery of local network connection patterns (in this case: triads of nodes), measuring their transitions during network evolution and present the preliminary results of this approach. We define the Triad Transition Matrix (TTM) containing the probabilities of transitions between triads, then we show how it can help to discover the dynamic patterns of network evolution. Also, we analyse the roles performed by different triads in the network evolution by the creation of triad transition graph built from the TTM, which allows us to characterize the tendencies of structural changes in the investigated network. The future applications of our approach are also proposed and discussed. © 2011 IEEE.
Kazienko, P, Bródka, P, Musial, K & Gaworecki, J 2010, 'Multi-layered social network creation based on bibliographic data', Proceedings - SocialCom 2010: 2nd IEEE International Conference on Social Computing, PASSAT 2010: 2nd IEEE International Conference on Privacy, Security, Risk and Trust, pp. 407-412.View/Download from: Publisher's site
A method for extraction of the multi-layered social network based on the data about human collaborative achievements, in particular scientific papers, is presented in the paper. The objects linking people form a hierarchy, which is flattened in the pre-processing stage. Only one level of the hierarchy remains together with new activities moved from its other levels. Separate layers of the multi-layered social network are created based on these pre-processed activities. © 2010 IEEE.
Kazienko, P, Brodka, P & Musial, K 2010, 'Individual neighbourhood exploration in complex multi-layered social network', Proceedings - 2010 IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology - Workshops, WI-IAT 2010, pp. 5-8.View/Download from: Publisher's site
Social networks can be extracted from different data about communication or common activities in organizations, companies or various Internet-based services. Different types of data processed may result in creation of separate layers in the complex multi-layered social network. Analysis of neighbourhoods of network members and their utilization to social group discovery appears to be an interesting and important research domain. Since there is no measure to evaluate structure of the neighbourhoods in the multi-layered social network, a new measure called cross layered multi-layered clustering coefficient (CLMCC) is proposed in the paper. It enables to analyse the density of mutual connections of neighbours that occur in at least a given number of layers in a social network. Additionally, experimental studies on real-world data are presented. © 2010 IEEE.
Bródka, P, Musial, K & Kazienko, P 2010, 'A method for group extraction in complex social networks', Communications in Computer and Information Science, pp. 238-247.View/Download from: Publisher's site
The extraction of social groups from social networks existing among employees in the company, its customers or users of various computer systems became one of the research areas of growing importance. Once we have discovered the groups, we can utilise them, in different kinds of recommender systems or in the analysis of the team structure and communication within a given population. The shortcomings of the existing methods for community discovery and lack of their applicability in multi-layered social networks were the inspiration to create a new group extraction method in complex multi-layered social networks. The main idea that stands behind this new concept is to utilise the modified version of a measure called by authors multi-layered clustering coefficient. © 2010 Springer-Verlag.
Juszczyszyn, K, Musiał, A, Musiał, K & Bródka, P 2010, 'Utilizing dynamic molecular modelling technique for predicting changes in complex social networks', Proceedings - 2010 IEEE/WIC/ACM International Conference on Web Intelligence and Intelligent Agent Technology - Workshops, WI-IAT 2010, pp. 1-4.View/Download from: Publisher's site
We present a method that utilises dynamic molecular modelling technique to track the changes within complex social network. The users forming a social network are interpreted as large sets of interacting particles. The data for the conducted research was obtained from e-mail communication within Enron company. The social network of employees was extracted and used to evaluate the methodology of social network dynamics modelling. © 2010 IEEE.
Brodka, P, Musial, K & Kazienko, P 2009, 'A performance of centrality calculation in social networks', CASoN 2009 - International Conference on Computational Aspects of Social Networks, pp. 24-31.View/Download from: Publisher's site
To analyze large social networks a lot of effort and resources are usually required. Network analysis offers many centrality measures that are successfully utilized in the process of investigating the social network characteristics. One of them is node position, which can be used to assess the importance of a given node within either the whole social network or the smaller subgroup. Three algorithms that can be utilized in the process of node position evaluation are presented in the paper: PIN Edges, PIN Nodes, and PIN hybrid. Also, different algorithms for indegree and outdegree prestige measures have been developed and tested. According to the experiments performed, the algorithms based onprocessing of edges are always faster than the others. © 2009 IEEE.
Brodka, P, Musial, K & Kazienko, P 2009, 'Efficiency of node position calculation in social networks', Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), pp. 455-463.View/Download from: Publisher's site
Social network analysis offers many measures, which are successfully utilized to describe the social network profile. One of them is node position, useful to assess the importance of a given node within both the whole network and its smaller subgroups. However, to analyze large social networks a lot of effort and resources are necessary. In this paper, some algorithms that can be utilized in the process of node position evaluation are presented and their efficiency is tested. In particular, three distinct algorithms were developed and compared: PIN Edges, PIN Nodes, and PIN hybrid. © 2009 Springer Berlin Heidelberg.
Musiał, K, Kazienko, P & Bródka, P 2009, 'User position measures in social networks', Proceedings of the 3rd Workshop on Social Network Mining and Analysis, SNA-KDD '09.View/Download from: Publisher's site
Network analysis offers many centrality measures that are successfully utilized in the process of investigating the social network profile. The most important and representative measures are presented in the paper. It includes indegree centrality, proximity prestige, rank prestige, node position, outdegree centrality, eccentrality, closeness centrality, and betweenes centrality. Both feature analysis and experimental comparative studies revealed the general profile of selected measures.
Juszczyszyn, K & Musiał, K 2009, 'Structural changes in an email-based social network', Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), pp. 40-49.View/Download from: Publisher's site
Different ways of detecting structural changes in email-based social networks are presented in the paper. A social network chosen for experiments was created on the basis of the Wroclaw University of Technology email server logs covering the period of 20 months. Structural parameters like degree centrality and prestige, clustering coefficients as well as betweenness and closeness centrality were computed for each of the consecutive months and their changes were analyzed. Our aim was to make an insight into dynamics of Internet-based social networks based on email service. It was found that the major changes in the structure of the network concern its local topology. Global indices like betweenness and closeness centrality remain relatively stable which also concerns the distribution of the local parameters such as degree centrality and prestige. However, the network size and local topology changes significantly which may be detected with motif analysis and visible changes in node clustering coefficients. © 2009 Springer Berlin Heidelberg.
Kazienko, P, Musiał, K & Zgrzywa, A 2009, 'Evaluation of node position based on email communication', Control and Cybernetics, pp. 67-86.
Rapid development of various kinds of social networks within the Internet enabled investigation of their properties and analyzing their structure. An interesting scientific problem in this domain is the assessment of the node position within the directed, weighted graph that represents the social network of email users. The new method of node position analysis, which takes into account both the node positions of the neighbors and the strength of connections between network nodes, is presented in the paper. The node position can be used to discover key network users, who are the most important in the population and who have potentially the greatest influence on others. The experiments carried out on two datasets enabled studying the main properties of the new measure.
Musial, K & Juszczyszyn, K 2009, 'Motif-based analysis of social position influence on interconnection patterns in complex social network', Proceedings - 2009 1st Asian Conference on Intelligent Information and Database Systems, ACIIDS 2009, pp. 34-39.View/Download from: Publisher's site
Motifs are small subgraphs showing statistically significant occurrence in given network. Motif analysis helps to insight into the local topology and functions of complex networks. The social position measure is interpreted as the importance of the node (user) within the network. We propose to fuse motif analysis with the social position assessment by colouring the nodes according to the measured position. As the distribution of discovered coloured motifs is utilized to mine the interconnection patterns between nodes, the results allow us to evaluate the influence of social position on the local topology of network connections. The experiment was carried out on the large social network derived from email communication. © 2009 IEEE.
Musiał, K & Juszczyszyn, K 2009, 'Properties of bridge nodes in social networks', Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), pp. 357-364.View/Download from: Publisher's site
The main goal of the paper is to describe the properties of the nodes within a social network that connect the peripheral nodes and peripheral groups with the rest of the network. These nodes are usually called bridging nodes or simply bridges. All the experiments are carried out on the real data, so-called Thurman network. First, the regular cliques, peripheral nodes and peripheral cliques from a network are extracted and then the bridging nodes identified. Afterwards for all nodes their characteristic features, such as social position and degree of nodes, are calculated. Finally, we try to find the correlation between nodes centrality and their degree and the fact if given node is a bridge or not. © 2009 Springer Berlin Heidelberg.
Musial, K, Juszczyszyn, K, Gabrys, B & Kazienko, P 2009, 'Patterns of interactions in complex social networks based on coloured motifs analysis', Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), pp. 607-614.View/Download from: Publisher's site
Coloured network motifs are small subgraphs that enable to discover and interpret the patterns of interaction within the complex networks. The analysis of three-nodes motifs where the colour of the node reflects its high - white node or low - black node centrality in the social network is presented in the paper. The importance of the vertices is assessed by utilizing two measures: degree prestige and degree centrality. The distribution of motifs in these two cases is compared to mine the interconnection patterns between nodes. The analysis is performed on the social network derived from email communication. © 2009 Springer Berlin Heidelberg.
Kazienko, P, Musiał, K & Kajdanowicz, T 2008, 'Profile of the social network in photo sharing systems', 14th Americas Conference on Information Systems, AMCIS 2008, pp. 2815-2826.
People, who interact, cooperate or share common activities within the photo sharing system can be seen as a multirelational social network. The results of their activities, i.e. tags, comments, references to favourites and others that semantically connect users through multimedia objects, i.e. pictures are the crucial component of the semantic web concept. Every online sharing system provides data that can be used for extraction of different kinds of relations grouped in layers in the multirelational social network. Layers and their profiles were identified and studied on two, spanned in time, snapshots of Flickr population for better understanding of social network structure complexity. Additionally, for each of the identified layers, a separate strength measure was proposed in the paper. The experiments on the Flickr photo sharing system revealed that users are inspired by both the semantic relationships between objects they operate on and social links they have to other users. Moreover, the density and affluence of the social network grows over course of time.
Musiał, K, Kazienko, P & Kajdanowicz, T 2008, 'Social recommendations within the multimedia sharing systems', Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), pp. 364-372.View/Download from: Publisher's site
The social recommender system that supports the creation of new relations between users in the multimedia sharing system is presented in the paper. To generate suggestions the new concept of the multirelational social network was introduced. It covers both direct as well as object-based relationships that reflect social and semantic links between users. The main goal of the new method is to create the personalized suggestions that are continuously adapted to users' needs depending on the personal weights assigned to each layer from the social network. The conducted experiments confirmed the usefulness of the proposed model. © 2008 Springer-Verlag Berlin Heidelberg.
Kazienko, P & Musiał, K 2008, 'Mining personal social features in the community of email users', Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), pp. 708-719.
The development of structure analysis that constitutes the core part of social network analysis is continuously supported by the rapid expansion of different kinds of social networks available in the Internet. The network analyzed in this paper is built based on the email communication between people. Exploiting the data about this communication some personal social features can be discovered, including personal position that means individual importance within the community. The evaluation of position of an individual is crucial for user ranking and extraction of key network members. The new method of personal importance analysis is presented in the paper. It takes into account the strength of relationships between network members, its dynamic as well as personal position of the nearest neighbours. The requirements for the commitment function that reflects the strength of the relationship are also specified. In order to validate the proposed method, the dataset containing Enron emails is utilized; first to build the virtual social network and afterwards to assess the position of the network members. © Springer-Verlag Berlin Heidelberg 2008.
Kazienko, P, Musiał, K & Juszczyszyn, K 2008, 'Recommendation of multimedia objects based on similarity of ontologies', Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), pp. 194-201.View/Download from: Publisher's site
A new framework for recommendation of multimedia objects based on individual ontologies is presented in the paper. The recommendation process takes into account similarities calculated both between objects' and users' ontologies that respect the social and semantic features existing in the system. The system was developed for the use inthe Flickr multimedia sharing system. © 2008 Springer-Verlag Berlin Heidelberg.
Juszczyszyn, K, Kazienko, P & Musiał, K 2008, 'Local topology of social network based on motif analysis', Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), pp. 97-105.View/Download from: Publisher's site
Network motifs - small subgraphs that reflect local topology can be used to discover general profile and properties of the network. Analysis of motifs for the large social networks derived from email communication is presented in the paper. The distribution of motifs in all analyzed real social networks is very similar one another and can be treated as the network fingerprint. This property is most distinctive for stronger human relationships. © 2008 Springer-Verlag Berlin Heidelberg.
Kazienko, P & Musiał, K 2007, 'Assessment of personal importance based on social networks', Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), pp. 529-539.
People that interact, cooperate or share common activities within information systems can be treated as a social network. The analysis of individual social standings appears to be a crucial element for the assessment of personal importance of each member within such weighted social network. The new measure of person significance - social position that depends on both the strength of relationships an individual maintains and social positions of all their acquaintances, together with its basic features and comparative experiments are presented in this paper. © Springer-Verlag Berlin Heidelberg 2007.
Kazienko, P & Musiał, K 2006, 'Social capital in online social networks', Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), pp. 417-424.
The problem of social capital in context of the online social networks is presented in the paper. Not only the specific elements, which characterize the single person and influence the individual's social capital like static social capital, activity component, and social position, but also the ways of stimulation of the social capital are described. © Springer-Verlag Berlin Heidelberg 2006.
My activities resulted in maintaining contacts with various research centres including the commercial companies’ research divisions of BT, Research and Engineering Centre, SAS Institute, Badoo, Phorm, and Affectv. The collaboration with Affectv resulted in joint Open Graph Initiative where we organised open challenges based on large social network data sets that come from 50 different social data providers all over the Europe. During my work at KCL I also collaborated and signed a substantial consultancy contract with Badoo, a company which owns a dating-focused social discovery website and has over 170 million active users in 180 countries. Another consultancy contract was signed with Phorm company and its goal was to explore machine learning approaches to analyse company data.
01/2014 – 03/2014 - Consultancy contract with Phorm. The main goal of the contract is to help company to understand in greater detail the mechanics of company's information system. The title of the project is: “Machine Learning approaches to analyse company data”. Income: £16,000.
01/2013 – 03/2014 - Consultancy agreement with Badoo Trading Limited aiming at understanding customer related data. The main goals of the contract were (i) to understand in greater detail the mechanisms of their email-based system, (ii) to develop a meaningful measure of user churn, and (iii) to discover the intentions of users. Income: £20,000.
Badoo was also involved in my EPSRC First Grant Application and was prepared to financially support the project (£60,000).
06/2013 – 01/2016 - Collaboration with Affectv company resulted in Open Graph Initiative in which the company opens up its aggregated social data for external world.
My collaboration with Affectv also resulted in their involvement in the application for the Centre for Doctoral Training in Data Science led by Bournemouth University (Prof. Bogdan Gabrys) in 2013.