Danilo Montesi

Spreading Processes in Multilayer Networks

Several systems can be modeled as sets of interconnected networks or networks with multiple types of connections, here generally called multilayer networks. Spreading processes such as information propagation among users of online social networks, or the diffusion of pathogens among individuals through their contact network, are fundamental phenomena occurring in these networks. However, while information diffusion in single networks has received considerable attention from various disciplines for over a decade, spreading processes in multilayer networks is still a young research area presenting many challenging research issues. In this paper, we review the main models, results and applications of multilayer spreading processes and discuss some promising research directions.

A Survey on Uncertainty Management in Data Integration

In the last few years, uncertainty management has come to be recognized as a fundamental aspect of data integration. It is now accepted that it may not be possible to remove uncertainty generated during data integration processes and that uncertainty in itself may represent a source of relevant information. Several issues, such as the aggregation of uncertain mappings and the querying of uncertain mediated schemata, have been addressed by applying well-known uncertainty management theories. However, several problems lie unresolved. This article sketches an initial picture of this highly active research area; it details existing works in the light of a homogeneous framework, and identifies and discusses the leading issues awaiting solutions.

Antiplatelet therapy and the outcome of subjects with intracranial injury: the Italian SIMEU study

IntroductionPre-injury antithrombotic therapy might influence the outcome of subjects withhead injuries and positive computed tomography (CT) scans. We aimed to determinethe potential risk of pre-injury antiplatelet drug use on short- and long-termoutcome of head injured subjects admitted to emergency departments (EDs) in Italyfor extended observation.MethodsA total of 1,558 adult subjects with mild, moderate and severe head injuryadmitted to Italian EDs were studied. In multivariable logistic regressionanalyses, the short-term outcome was assessed by an evaluation of head CT scan at6 to 24 hours after trauma and the long-term outcome by the Glasgow outcome scale(GOS) at six months.ResultsHead CT scan comparisons showed that 201 subjects (12.9%) worsened. The risk ofworsening was increased two fold by the use of antiplatelet drugs (106, 19.7%treated versus 95, 9.3% untreated; relative risk (RR) 2.09, 95% CI 1.63 to 2.71).The risk was particularly high in subjects on clopidogrel (RR 5.76, 95% CI 3.88 to8.54), independent of the association with aspirin. By logistic regression, 5 of14 items were independently associated with worsening (Glasgow coma scale (GCS),Marshall category, antiplatelet therapy, intraventricular hemorrhage, number oflesions). After six months, only 4 of 14 items were predictors of unfavorableoutcome (GOS 1 to 3) (GCS score, Marshall category, age in decades, intracerebralhemorrhage/contusion). The risk increased by 50% in the group treated withantiplatelet therapy (RR 1.58, 95% CI 1.28 to 1.95; P < 0.001).ConclusionsAntithrombotic therapy (in particular clopidogrel) is a risk factor for bothshort-term and long-term unfavorable outcome in subjects with head injury,increasing the risk of progression and death, permanent vegetative state andsevere disability.

BPMN: how much does it cost? an incremental approach

In this paper we propose some extensions of the businessprocess modeling notation (BPMN) to be able to evaluate the overallcost of business process diagrams. The BPMN is very expressive, and ageneral treatment of this problem is very complex. Therefore, it seemsreasonable to define classes of business process diagrams capturing realprocesses and to develop efficient analysis methods for these classes. Inthe paper we define some relevant subsets of the BPMN, extend themwith the concept of cost, and provide computational models for eachclass, in most cases reducing them to existing problems for which efficientsolutions already exist.

Information Propagation Analysis in a Social Network Site

One of the most interesting and still not completely understood phenomena happening in Social Network Sites is their ability to spread (or not) units of information which may aggregate to form large distributed conversations. In this paper we present the result of an empirical study on a Large Social Database (LSD) aimed at measuring the factors enabling information spreading in Social Network Sites.

Conversation retrieval for microblogging sites

In this article we introduce a novel search paradigm for microblogging sites resulting from the intersection of Information Retrieval and Social Network Analysis (SNA). This approach is based on a formal model of on-line conversations and a set of ranking measures including SNA centrality metrics, time-related conversational metrics and other specific features of current microblogging sites. The ranking approach has been compared to other methods and tested on two well known social network sites (Twitter and Friendfeed) showing that the inclusion of SNA metrics in the ranking function and the usage of a model of conversation can improve the results of search tasks.

Friendfeed Breaking News: Death of a Public Figure

Microblogging sites allow users to post short messages online, offering a reliable way to communicate and to spread information quickly and efficiently; moreover, they can host complex conversational activities. The objectives of this paper are to model how breaking news circulate in a microblogging network, to identify relevant patterns of news propagation and to increase our understanding of the underlying sociological motivations. This study has been conducted on a real social database collected from a well known microblogging site by observing the reactions of its users to a relevant public event.

A unified approach to structured and XML data modeling and manipulation

In this paper we propose an approach to defining logical database models, based on the instantiation of a general abstract model, and discuss its application to the management of mixed XML/relational data. Our abstract model is equipped with a parametric query algebra and relational-like algebraic equivalences, that do not have to be redefined when new models are generated. We present an instantiation of our model that unifies the main approaches to represent and manipulate relational and XML data, and in particular SQL, SQL/XML, XQuery, and Oracles XML Type. Additionally, our algebra can represent queries not expressible by other algebras taken from the literature. Among others, it can represent nested XQuery expressions with no constraints on nesting and on node constructors.

Missing Data in Multiplex Networks: A Preliminary Study

A basic problem in the analysis of social networks is missing data. When a network model does not accurately capture all the actors or relationships in the social system under study, measures computed on the network and ultimately the final outcomes of the analysis can be severely distorted. For this reason, researchers in social network analysis have characterised the impact of different types of missing data on existing network measures. Recently a lot of attention has been devoted to the study of multiple-network systems, e.g., Multiplex networks. In these systems missing data has an even more significant impact on the outcomes of the analyses. However, to the best of our knowledge, no study has focused on this problem yet. This work is a first step in the direction of understanding the impact of missing data in multiple networks. We first discuss the main reasons for missingness in these systems, then we explore the relation between various types of missing information and their effect on network properties. We provide initial experimental evidence based on both real and synthetic data.

Conversation retrieval from twitter

The process of retrieving conversations from social network sites differs from traditional Web information retrieval because it involves human communication aspects, like the degree of interest in the conversation explicitly or implicitly expressed by the interacting people and their influence/popularity. Our demo allows users to include these aspects into the search process. The system allows the retrieval of millions of conversations generated on the popular Twitter social network site, and in particular conversations about trending topics.