1,720,989 research outputs found
Social Network to analyse the relationship between ‘victim-author’ and ‘motivation’ of violence against women in Italy.
No full textThe paper aims to analyse the phenomenon of Violence against women in the Italian context during 2020. It proposes to study the relationship between ‘victim-author’ and ‘motivation’ in femicides committed in domestic environment. By means of the properties of the Social Network Analysis on bimodal data, the study detected main actors and motivations that generated the homicides with female victims. At the same time, the structural relationships allowed to investigate the existence of motivations that better characterized the action of the various actors. The bipartite graph visualization and centrality scores calculated have demonstrated the effectiveness of the methodology for the pursued objectives
Data Quality and Violence Against Women: The Causes and Actors of Femicide
No full textThe paper examines domestic ’femicide’ in Italy. Under an exploratory statistical approach, we investigated: (1) difficulties and strategies for reconstructing a historical dataset on fam- ily crimes for studies over time; (2) the main causes of family femicides; and (3) groups of actors driven by the same motivations interpreted as patterns of criminal behavior. First, we integrated and systematised data from official sources to guarantee comparison over time; second, we used Social Network Analysis properties to study the relationships between ’motivations’ and ’victim-perpetrator’; and third, we applied and compared com- munity detection algorithms to the linkages between ’actors’ and ’motivations’ to detect groups of criminal behavior. From 2015 to 2020 in Italy, the cohabitant was the major fam- ily murderer, but in 2020, passion motivation also surfaced. Mental problems connected to parents-children and cohabitants, jealousy of ex-partners or rivals, and economic issues for blood relations were observed in 2015. Psychopathologies and money characterised par- ents-children in 2020, while passion and disagreements caused cohabitants or ex-partners
Da informativa ad ostile: analisi e profilazione della comunicazione dei leader al tempo del Covid-19
No full textThe contribution highlights the transformations in the communication strategies of Italian political leaders between the first and second waves of the Covid-19 pandemic, in a broader context of changes in political strategies and contingent political assets. Specifically, the study focuses on the political language adopted on Twitter by seven figures who, due to their political role and popularity, dominate the national political scene during the monitoring period: Silvio Berlusconi, Giuseppe Conte, Luigi Di Maio, Giorgia Meloni, Matteo Renzi, Matteo Salvini and Nicola Zingaretti. Through the combination of the Latent Dirichlet Allocation topic modeling algorithm and multidimensional analysis techniques, the document analyzes and compares the language used in tweets by political actors during the first pandemic wave (March-May 2020) and the second one (September - December 2020). The factorial projections based on the association between leaders and the keywords of political topics allowed to define four types of political communication: informative, propagandistic, demagogic and hostile. The main contribution of this work consists in a combination of statistical techniques to profiling and studying changes in the communication strategies of leaders. The results show that in the first wave - characterized by a perception of disorientation and suspension of judgment - the exponents of the government majority use a communication aimed at providing information and limiting conflicts on the political front, while the opposition actors adopt a hostile communication aimed at bringing their issues back to the center of public debate. In the second wave - when the state of crisis is assimilated, metabolized and normalized within society - the dynamics of fragmentation of the public sphere, typical of the confrontation in the digital media arena, become evident again and the pandemic becomes a new divisive and polarizing within the political debate, in which each party returns to entrench itself on its own ideological positions. Therefore, the main change found is the shift from an informative political communication, based on rational speeches, to a hostile one, based on polemical language aimed at delegitimizing the acts of the government and leveraging the affectivity of the public
The Revolution of AI in Healthcare Between Diagnosis and Treatment
No full textThe complexity and volume of data in healthcare entail that artificial intelligence (AI) and associated technologies are becoming an essential component of life sciences. The scientific literature explores the benefits, contexts of application, ethical implications, and future devel- opments. In order to 1) identify the topics characterizing the literature on AI in the clinical domain, 2) detect semantic categories, and 3) val- idate them, a methodological approach based on the combination of natural language processing and machine learning for classification was performed. Two main semantic categories were identified: diagnostics and treatment, which are used to manually annotate each document. Finally, we tested our semantic classification through machine learning. The findings suggest clear differences between the two categories, mainly based on AI-assisted meta-analyses and clinical decision support systems, with just a quota of scientific papers encompassing both semantic pil- lars. This proportion of documents is pivotal to changing the semantic classification
Third Mission & VQR 2015-2019: a bigram’s story
No full textThird Mission VQR exercise for 2015–2019 emphasised a competitive landscape for state and non-state universities. Each institution has its own history, mission and strategic perspective, each with its own description of TM contained inside the strategic plan. The goal of this article is to investigate the content of strategic plans made by Italian universities through 2022 and their relationship with VQR scores using two research questions. The replies were obtained by network textual analysis and correspondence analysis based on bigrams
Healthcare and AI frontiers: thematic insights and domains categorization
No full textThe integration of Artificial Intelligence (AI) has become essential in life sciences research. Scientific literature actively explores AI technologies, their strengths, ethical considerations, application fields, and future developments. In this framework, this paper analyzes academic literature on AI in healthcare, pursuing four research directions: a) providing a comprehensive overview of AI advancements in healthcare; b) identifying key themes related to technologies and application areas; c) classifying scientific literature based on the semantic labels from point b); d) determining the sustainability implications of AI in healthcare. A corpus of abstracts published between 2000 and 2023 in the Web of Science database was analyzed. A methodological framework using natural language processing (NLP) was implemented through a topic model in embedding spaces. The trend shifted from image analysis for diagnosis and treatment in the early 2000 s to predictive models for fields such as ophthalmology, internal medicine, and oncology. New technological systems have emerged to support computer-assisted tools, including chatbots, telemedicine, and AI-driven administrative tasks. Sustainability implications were detected across technological, environmental, and social dimensions. Machine learning and deep learning optimize hospital management, diagnostics, and resource allocation, reducing costs and waste. Robotics enhances surgical precision and patient care, while NLP improves data analysis and decision-making. Environmentally, AI-driven telemedicine reduces the ecological footprint, supporting climate goals. Socially, AI fosters healthcare equity by personalizing treatments and addressing ethical concerns related to data transparency. The categorization of the literature was evaluated by comparing two machine learning models, Support Vector Machine and Random Forest, providing insights into the current and future directions of AI in healthcare
A Proposal for Cross-language Analysis: violence against women and the Web.
No full textAim of the paper is investigating the mood on the Web with respect to one of the most relevant Human Rights violation, without any geographic distinction: the violence against women. While the literature that studies the phenomenon is rapidly growing, the action field is still fragile and the question marks are about the relationship between the public opinion and the contextual factors. In a first look at the phenomenon, we aim at mapping gender violence on the Web, in a Big Data perspective. The peculiar problem we deal with consists in analysing short documents (tweets) written in six European different languages, in the occasion of a common event: the International Day for the Elimination of Violence against Women, 25 November 2017. For our statistical analysis, we choose a multi-linguistic, cross-national perspective. The basic idea is that there are some common structures, language independent ("concepts"), which are declined in the different national natural language expressions ("terms"). Investigating those structure (e.g. factors of lexical correspondence analyses separately performed on the different collections), enables a double level analysis trying to understand and visualise national peculiarities and communalities. The statistical tool is given by Procrustes rotations. Keywords: Big Data, Text Mining, Cross-national study, Procrustes rotation
L'impatto del Covid-19 sull'opinione pubblica: una strategia di analisi per lo studio della comunicazione su Twitter
No full textEmotion recognition in Italian political language to predict positionings and crises government.
Full text linkThe paper aims to analyze the political language adopted on Twitter by the main Italian parties’ leaders during the first two waves of Covid-19 pandemic.
A two-step model based on sentiment emotion recognition (ER) and Correspondence analysis detected which emotions characterized the political language and which changes happened between the two waves. The results showed the use of a language with a strong emotional weight for some political actors as opposed to others who used a neutral register of political language in both waves. The comparison between two waves denoted a shift from anger to sadness and fear for Meloni and a moving away Salvini by predicting through ER the rift of the right-wing
- …
