1,721,531 research outputs found
Handling uncertainty in clustering art-exhibition visiting styles
No full textUncertainty is one of the most critical aspects that affect the quality of Big Data management and mining methods. Clustering uncertain data has traditionally focused on data coming from location- based services, sensor networks, or error-prone laboratory experiments. In this work we study for the first time the impact of clustering uncertain data on a novel context consisting in visiting styles in an art exhibition. We consider a dataset derived from the interaction of visitors of a museum with a complex Internet of Things (IoT) framework. We model this data as a set of uncertain objects, and cluster them by employing the well-established UK-medoids algorithm. Results show that clustering accuracy is positively impacted when data uncertainty is taken into account. © ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering 2017
Influence of some parameters on visiting style classification in a cultural heritage case study
No full textA smart system for a cultural exhibition, generally, has the ability to infer interests of users and to track the propagation of the information into the event. We are interested in analysing and studying the visiting styles of users in a real cultural heritage exhibition, named The Beauty or the Truth. Starting from data that was collected during this exhibition, the interaction between visitors and artworks and the influences of the available technology on their behaviours are studied. Finally, we analyse how the tuning of some parameters on a classification strategy influences the users’ visiting styles. The obtained results have revealed interesting issues also to understand hidden aspects in the data and unattended in the analysis. © Springer International Publishing Switzerland 2016
Influence analysis in business social media
No full textIn this paper, we describe a novel data model for particular online business social networks such as Tripadvisor and Yelp: we also define a greedy influence maximization algorithm to determine the most influential users on the base of proper influence patterns. The result of such analysis is then combined with some economic data in order to propose a set of possible financial strategies for business objects. Finally, a case study and some preliminary and interesting results are presented for the Yelp dataset
An information-theoretic approach to hierarchical clustering of uncertain data
No full textUncertain data clustering has become central in mining data whose observed representation is naturally affected by imprecision, staling, or randomness that is implicit when storing this data from real-word sources. Most existing methods for uncertain data clustering follow a partitional or a density-based clustering approach, whereas little research has been devoted to the hierarchical clustering paradigm. In this work, we push forward research in hierarchical clustering of uncertain data by introducing a well-founded solution to the problem via an information-theoretic approach, following the initial idea described in our earlier work [26]. We propose a prototype-based agglomerative hierarchical clustering method, dubbed U-AHC, which employs a new uncertain linkage criterion for cluster merging. This criterion enables the comparison of (sets of) uncertain objects based on information-theoretic as well as expected-distance measures. To assess our proposal, we have conducted a comparative evaluation with state-of-the-art algorithms for clustering uncertain objects, on both benchmark and real datasets. We also compare with two basic definitions of agglomerative hierarchical clustering that are treated as baseline methods in terms of accuracy and efficiency of the clustering results, respectively. Main experimental findings reveal that U-AHC generally outperforms competing methods in accuracy and, from an efficiency viewpoint, is comparable to the fastest baseline version of agglomerative hierarchical clustering. © 2017 Elsevier Inc
Visiting Styles in an Art Exhibition Supported by a Digital Fruition System
No full textWe investigate the user dynamics related to the interaction with artworks in an exhibition. In a first step, we characterize visitors in a cultural heritage scenario and after, we study how these interact with available technologies. Accordingly with the fact that the technology plays a crucial role in supporting spectators and enhancing their experiences, the starting point of this research is the analysis of real data coming from visitors of the art exhibition named The Beauty or the Truth that was located in Naples, Italy. The event was equipped with several technological tools arranged within the halls of the exhibition, with the aim to create a novel metaphor that stimulates the user enjoyment and the knowledge diffusion. The collected log files from a suitable expert software system are used in a flexible framework in order to analyse how the supporting pervasive technology influence and modify behaviours and visiting styles. Finally, we carried out some experiments to exploit the clustering facilities for finding groups that reflect visiting styles. The obtained results have revealed interesting issues also to understand hidden aspects in the data and unattended in the analysis. © 2015 IEEE
Be certain of how-to before mining uncertain data
No full textThe purpose of this technical note is to introduce the problems of similarity detection and summarization in uncertain data. We provide the essential arguments that make the problems relevant to the data-mining and machine-learning community, stating major issues and summarizing our contributions in the field. Further challenges and directions of research are also issued. © 2014 Springer-Verlag
A Numerical Approach for Assigning a Reputation to Users of an IoT Framework
No full textNowadays, in the Internet of Things (IoT) society, the massive use of technological devices available to the people makes possible to collect a lot of data describing tastes, choices and behaviours related to the users of services and tools. These information can be rearranged and interpreted in order to obtain a rating (i.e., evaluation) of the subjects (i.e., users) interacting with specific objects (i.e., items). Generally, reputation systems are widely used to provide ratings to products, services, companies, digital contents and people. Here, we focus on this issue, adopting a Collaborative Reputation System (CRS) to evaluate the visitors' behaviour in a real cultural event. The results obtained, compared with those obtained by other methods (i.e., classification), have confirmed the reliability and the usefulness of CRSes for deeply understand dynamics related to visiting styles
Collaborative reputation systems in a cultural heritage scenario
No full textIn the last decade, algorithms for reputation systems are been widely developed in order to achieve correct ratings for products, services, companies, digital contents and people. We start from a comprehensive mathematical model for Collaborative Reputation Systems (CRSes), present in the literature and formally defined as a recurrence relation that generates a sequence of trust matrices, from which the reputation of the items and the raters can be derived. Even though this model can be applied to several scenarios, the focus of this work is related to its application in a real case, that is a cultural event scenario. More in detail, in cultural heritage environment, the data collected in an event represent the basic knowledge to be inferred. The main idea is to correctly use the available technology and data to give a reliable rate (reputation) for both visitors and artworks. These rates will be very useful to classify the visiting style of the visitors and to fix the artworks that have most attracted visitors
A probabilistic approach for financial IoT data
No full textThe extraction of information from the Internet of Things (IoT) plays a fundamental role in many research fields. In this work we focus our attention on financial data, used to describe self-financing portfolios in a complete market. Here, the absence of the arbitrage principle, the existence and the uniqueness of no arbitrage price are valid. With these hypotheses we can resort to the Black-Scholes model in order to determine the expression of no arbitrage price. In this model, frictional costs are avoided. Moreover, selling and buying of every amount of the assets and short sellings are allowed. In other words, traders can sell amount of assets even if they do not own them. Finally, this model is composed by a risk-free and a Geometric Brownian motion risk assets. © 2016, CEUR-WS. All rights reserved
MMVar: Clustering uncertain objects via minimization of the variance of cluster mixture models
No full textA major issue in clustering uncertain objects is related to the poor efficiency of existing algorithms, which is mainly due to expensive computation of the distance between uncertain objects. This paper discusses how we addressed this issue through an original formulation of the problem of clustering uncertain objects based on the minimization of the variance of the mixture models that represent the clusters to be discovered. The proposed partitional clustering method, named MMVar, features high efficiency since it does not need to employ any distance measure for uncertain objects. Experiments have shown that MMVar turned out to be faster than prominent state-of-the-art algorithms for clustering uncertain objects, while achieving better average accuracy in terms of both external and internal cluster validity criteria
- …
