1,720,995 research outputs found
Fake Account Identification in Social Networks
No full textNowadays, the human influence often depends on the number of followers that an individual has in his/her own social media context. To this end, the presence of fake accounts is one of the most relevant problems and can potentially have a big impact on many real life and business activities. Fake followers are dangerous for social platforms, since they may alter concepts like popularity and influence, which might yield a strong impact on economy, politics, and society. Thus, it is necessary to devise new methodologies enabling the possibility to identify and characterize fake accounts. This work presents a novel technique to discriminate real accounts on social networks from fake ones. The technique exploits knowledge automatically extracted from big data to characterize typical patterns of fake accounts. We empirically evaluated the proposed technique on the Twitter social network, and achieved significant results in terms of discrimination capabilities
Lattice-based Discovery of Hybrid Relaxed Functional Dependencies
No full textRelaxed functional dependencies (rfds) are properties expressing important relationships among data. Thanks to the introduction of approximations in data comparison and/or validity, they can capture constraints useful for several purposes, such as the identi fication of data inconsistencies or patterns of semantically related data. Nevertheless, rfds can provide benefits only if they can be automatically discovered from data. In this discussion paper we present an rfd discovery algorithm relying on a lattice structured search space, and a new candidate rfd validation method. An experimental evaluation demonstrates the discovery performances of the proposed algorithm on real datasets
Efficient validation of functional dependencies during incremental discovery
No full textThe discovery of functional dependencies (FDs) from data is facing novel challenges also due to the necessity of monitoring datasets that evolves over time. In these scenarios, incremental FD discovery algorithms have to efficiently verify which of the previously discovered FDs still hold on the updated dataset, and also infer new valid FDs. This requires the definition of search strategies and validation methods able to analyze only the portion of the dataset affected by new changes. In this paper we propose a new validation method, which can be used in combination with different search strategies, that exploits regular expressions and compressed data structures to efficiently verify whether a candidate FD holds on an updated version of the input dataset. Experimental results demonstrate the effectiveness of the proposed method on real-world datasets adapted for incremental scenarios, also compared with a baseline incremental FD discovery algorithm
Visual ECG analysis in real-world scenarios
No full textCardiac arrhythmia is an alteration of the heart rhythm, for which the heartbeat is irregular. Based on the severity of this condition, an arrhythmia could represent a serious danger for a patient. An ECG is a graphic representation of an heart rhythm, which provides an overview of heart's conditions over a specific time interval. ECG signal analysis is entrusted to trained clinicians, although complex and frantic environments, such as emergency settings, can make hard to delegate continuous monitoring to the medical personnel. In such scenarios, an automatic detection methodology could provide crucial support in promptly alerting clinicians towards a potential degeneration of a patient's conditions. To this end, we propose a heartbeat classification module capable of capturing the semantics of visual information of ECG signals provided by video frames. The module relies on feature extraction techniques derived from video projected images resulting in ECG data, which are then classified by means of deep-learning models. It can be used to support the early detection of some arrhythmia in critical contexts, such as emergency rooms. We show how the proposed module can be used to support clinicians in this context, and discuss an experimental evaluation performed over ground-truth datasets
RYAN: A tool for explaining and visually analyzing the evolution of Relaxed Functional Dependencies
No full textDiscovering relaxed functional dependencies based on multi-attribute dominance [extended abstract]
No full textTo assess the quality of data it is useful to extract properties and relationships among them. However, exceptions and approximations need be considered in real-world settings. To this end, relaxed FDs (REDS) are data dependencies accounting for both exceptions and similarities on data, but their discovery is an extremely complex problem, also due to the necessity of specifying similarity and validity thresholds. The RFD discovery algorithm presented in this paper exploits the concept of dominance to automatically derive similarity thresholds. The discovery performances and the effectiveness of the proposed algorithm are assessed through a comparative evaluation with state-of-art approaches
Decentralized and Incremental Discovery of Relaxed Functional Dependencies Using Bitwise Similarity
No full textOver the past decade, there have been numerous extensions to the definition of Functional Dependency (FD), culminating in the introduction of Relaxed Functional Dependency (RFD), offering more flexible constraints compared to traditional FDs. This increased flexibility makes RFDs well-suited for exploring and profiling data in datasets with lower data quality. However, efficiently identifying RFDs within dynamic data sources presents a significant challenge, as it requires processing an entire dataset from scratch whenever modifications occur. To tackle this problem, incremental discovery algorithms have been defined, but they often suffer when the frequency and the size of batches of updates increase. This paper presents a new algorithm, namely D-INDIBITS, relying on a new decentralized architecture to balance the workload that drives the incremental discovery process of INDIBITS, which is based on bitwise operators for computing attribute similarities. Experiments demonstrate DINDIBITS's effectiveness compared to FD and RFD discovery algorithms on both static and dynamic real-world data. With batches of modifications of sizes 10k and 100k, D-INDIBITS is capable of updating the set of RFDs in a few seconds, whereas all other approaches often employ more than 3 hours
Analyzing the worldwide perception of the Russia-Ukraine conflict through Twitter
No full textIn this paper, we analyze the worldwide perception of the Russia-Ukraine conflict (RU conflict for short) on the Twitter platform. The study involved collecting over 17 million tweets written in 63 different languages and conducting a multi-language sentiment analysis, as well as an analysis of their geographical distribution and verification of their temporal relationship to daily events. Additionally, the study focused on analyzing the accounts producing pro-conflict tweets to evaluate the possible presence of bots. The results of the analysis showed that the war had a significant global impact on Twitter, with the volume of tweets increasing as the war’s threats materialized. There was a strong correlation between the succession of events, the volume of tweets, and the prevalence of a specific sentiment. Most tweets had a negative sentiment, while tweets with positive sentiment mainly contained support and hope for people directly involved in the conflict. Moreover, a bot detection analysis performed on the collected tweets revealed the presence of many accounts spreading tweets including pro-conflict hashtags that cannot be identified as real users. Overall, this study sheds light on the importance of social media in shaping public opinion during conflicts and highlights the need for reliable methods to detect bots
Dependency Visualization in Data Stream Profiling
No full textData stream profiling concerns the automatic extraction of metadata from a data stream, without having the possibility to store it. Among the metadata of interest, functional dependencies (FDs), and their extensions relaxed functional dependencies (RFDs), represent an important semantic property of data. Nowadays, there are many algorithms for automatically discovering them from static datasets, and some are being proposed for data streams. However, one of the main problems is that the stream nature of data requires a different paradigm of monitoring, since the “big” number of (R)FDs that might hold on a given dataset continuously change as new data are read from the stream. In this paper, we present a tool for visualizing RFDs discovered from a data stream. The tool permits to explore results for different types of RFDs, and uses quantitative measures to monitor how discovery results evolve. Moreover, the tool enables the comparison among RFDs discovered across several executions, also proving visual manipulation operators to dynamically compose and filter results. A user study has been conducted to assess the effectiveness of the proposed visualization tool
- …
