https://www.um.edu.mt/library/oar/handle/123456789/5581 2026-04-27T02:25:43Z https://www.um.edu.mt/library/oar/handle/123456789/144778 Title: How social imbalance and governance quality shape policy directives for energy transition in the OECD countries? Authors: Sinha, Avik; Bekiros, Stelios; Hussain, Nazim; Nguyen, Duc Khuong; Khan, Sana Akbar Abstract: In line with the COP26 Summit objectives, this paper develops a policy framework to achieve energy transition by considering the social imbalances and regulatory effectiveness. A new energy transition index is proposed. It is an output-side indicator based on the energy ladder hypothesis. This index enables to apprehend the energy transition scenario in any country by capturing (a) the transition to a cleaner energy source, and (b) the transition to more energy efficient sources. Using the two-step System GMM approach and data for 37 OECD countries over the 2000–2019 period, the dynamic and extreme marginal impacts of energy transition drivers with respect to estimates of the model parameters are analyzed. The results show that the social imbalance dampens the positive impacts of energy transition drivers, whereas governance quality helps in augmenting those impacts. The outcomes, drawn from a scenario-based policy design approach, are particularly helpful in advancing potential policy discourse. They have important practical implications for the development of the SDG-oriented policy framework, with special focus on the attainment of the SDG 7 and 13. 2023-01-01T00:00:00Z https://www.um.edu.mt/library/oar/handle/123456789/144573 Title: A comparative assessment of machine learning methods for predicting housing prices using Bayesian optimization Authors: Lahmiri, Salim; Bekiros, Stelios; Avdoulas, Christos Abstract: The valuation of house prices is drawing noteworthy attention due to worldwide financial and real estate crises in the last decade. Therefore, there is an immediate need to design more effective predictive systems of house prices. Indeed, investors, creditors, and governments are all interested in such predictive systems to improve their buying and lending decisions and activities. This study explores the application of artificial intelligence, machine learning, and nonlinear statistical models to house price prediction problems. In that order, we use boosting ensemble regression trees, support vector regression, and Gaussian process regression. Bayesian optimization is implemented in a ten-fold cross-validation framework to determine their respective optimal kernels and parameter values. Four performance metrics are used to evaluate the prediction ability of each predictive system. The experimental results showed that boosting ensemble regression trees performed the best, followed by Gaussian process regression and support vector regression. In addition, all three aforementioned predictive systems outperformed artificial neural networks and multi-variate regression employed in recent work on the same data set. Under this perspective, it is concluded that boosting ensemble regression trees are clear candidates to be considered for operational house price prediction in Taiwan. 2023-01-01T00:00:00Z https://www.um.edu.mt/library/oar/handle/123456789/141836 Title: Cutting edge recruitment best practices : a REA study Abstract: From a structural functionalistic perspective, this dissertation sought to understand how best to align the processes of formal education, and the processes of recruitment, such that both develop in syntony, rather than in disharmony. Two knowledge synthesis tools were employed. In the first part (Chapters 1 and 2 specifically, together with their related appendices), a traditional literature review was undertaken, for the purpose of understanding how modern economies are trending to begin with, and thus extrapolate the likely economic scenarios of the future. The evidence relied upon in this review consisted of expert opinion, and up-to-date statistical information. In the second part (Chapters 3 and 4 specifically, together with their related appendices), a Rapid Evidence Assessment (REA) of influential academic literature was undertaken. The REA specifically analysed thirty-one sources, specifically consisting of sixteen meta-analyses, two systematic reviews, four cross-sectional studies, and nine traditional literature reviews, that spanned almost half a century of research, from 1973 to 2022 (please note Appendix 2 specifically). The systematic manner (please note Chapter 3) by which these thirty-one sources were identified and selected, was novel, and the results appear to have been promising and fruitful. Additionally, apart from the systematic overview of the said academic literature, this REA made an original discovery concerning the criteria of job performance, with the systematically collected evidence clearly suggesting that there is a hierarchical order to these criteria, meaning that some of these criteria consistently yield higher validities than others, irrespective of the type of predictor or occupational category (please note RQ3’s Executive Summary). The key factual findings of this dissertation have been listed in Appendices 32 and 33, for the reader’s convenience. Description: M.A.(Melit.) 2025-01-01T00:00:00Z https://www.um.edu.mt/library/oar/handle/123456789/141835 Title: The challenges of deepfake technology on the decision-making processes within law enforcement : a study within the EU landscape Abstract: This study aims to scrutinize the challenges posed by deepfake technology in the decision-making processes of law enforcement, both locally and across Europe, categorizing them into three key areas:  Knowledge and Awareness: Assessing law enforcement officers’ familiarity with deepfake technologies, associated crimes, and the operational challenges encountered.  Evidence Management: Identifying difficulties in detecting deepfake content, preserving its evidential integrity, and ensuring its admissibility.  Legal and Procedural Challenges: Analyzing legislative frameworks and courtroom practices regarding deepfake evidence, and current initiatives to mitigate misuse. A qualitative research approach was employed, utilizing a structured questionnaire distributed through intermediaries to law enforcement agencies across the EU, supplemented by secondary data and case studies to enhance validity through triangulation. Twelve officers responded (4 from Malta and 8 from other EU countries), all with direct or indirect experience of deepfakes in criminal investigations. There was a broad consensus among respondents on a significant lack of preparedness in addressing deepfake threats, both in terms of detection capabilities and evidentiary handling. Officers reported substantial difficulties in reliably identifying manipulated media, preserving its chain of custody, and navigating inconsistent legal standards across jurisdictions. Detection challenges are compounded by rapid advances in synthetic media quality, often outpacing available forensic tools. In contrast to findings from EU respondents, studies from the United States and Australia show a slightly higher level of operational preparedness, often attributed to earlier adoption of digital forensic techniques and a more rapid development of AI-focused legal scholarship. However, even in these jurisdictions, significant concerns remain about evidentiary reliability and procedural fairness when presenting AI-generated content. Deepfake evidence presents unique admissibility challenges, such as establishing authenticity, reliability, and the absence of tampering. Courts are struggling to develop consistent standards, and current digital evidence frameworks often lack explicit provisions for synthetic media. There is a growing debate over whether new evidentiary rules are needed or if existing frameworks (such as chain of custody, expert testimony, and metadata analysis) can adapt adequately. On the ground, officers report that gaps in regulation and procedural clarity lead to uncertainties in evidence collection and presentation, delaying investigations or causing reliance on expert witnesses to establish basic authenticity. Prosecutors express concern over juror perceptions of manipulated media and the risk of undermining trust in legitimate evidence. These issues often translate into higher costs, longer case preparation times, and difficulties in achieving convictions. Description: M.A.(Melit.) 2025-01-01T00:00:00Z