Birmingham City University Open Access Repository
Not a member yet
8823 research outputs found
Sort by
Deep Behavioral Analysis of Machine Learning Algorithms Against Data Poisoning
No full textPoisoning attacks represent one of the most common and practical adversarial attempts on machine learning systems. In this paper, we have conducted a deep behavioural analysis of six machine learning (ML) algorithms, analyzing poisoning impact and correlation between poisoning levels and classification accuracy. Adopting an empirical approach, we highlight practical feasibility of data poisoning, comprehensively analyzing factors of individual algorithms affected by poisoning. We used public datasets (UNSW-NB15, BotDroid, CTU13, and CIC-IDS-2017) and varying poisoning levels (5% - 25%) to conduct rigorous analysis across different settings. In particular, we analyzed the accuracy, precision, recall, f1-score, false positive rate and ROC of the chosen algorithms. Further, we conducted a sensitivity analysis of each algorithm to understand the impact of poisoning on its performance and characteristics underpinning its susceptibility against data poisoning attacks. Our analysis shows that, for 15% poisoning of UNSW NB15 dataset, the accuracy of Decision Tree (DT) decreases by 15.04% with an increase of 14.85% in false positive rate. Further, with 25% poisoning of BotDroid dataset, accuracy of K-nearest neighbours (KNN) decreases by 15.48%. On the other hand, Random Forest (RF) is comparatively more resilient against poisoned training data with a decrease of 8.5% in accuracy with 15% poisoning of UNSW-NB15 dataset and 5.2% for BotDroid dataset. Our results highlight that 10%-15% of dataset poisoning is the most effective poisoning rate, significantly disrupting classifiers without introducing overfitting, whereas 25% is detectable because of high performance degradation and overfitting algorithms. Our analysis also helps understand how asymmetric features and noise affect the impact of data poisoning on machine learning classifiers. Our experimentation and analysis are publicly available at: https://github.com/AnumAtique/Behavioural-Analaysis-of Poisoned-ML
Eating in response to emotions: Alexithymia, emotional eating, and associated psychological mechanisms
No full textThe overarching objective was to elucidate the relationship between alexithymia and eating in response to emotions. First, a systematic review synthesised the findings of nine eligible articles, providing preliminary evidence for a positive association between alexithymia and self-reported emotional eating. As the Dutch Eating Behaviour Questionnaire (DEBQ-EE) was the subjective emotional eating measure most frequently used by previous research, it became the subject of an exploratory ‘think aloud’ study. This study audio-recorded participants’ spoken aloud thoughts as they completed the DEBQ-EE online. Two cross-sectional studies were conducted to further explore alexithymia and emotional eating using other self-report measures (Emotional Eating Scale [EES] and Salzburg Emotional Eating Scale [SEES]) and identify mechanisms for potential intervention targets. Findings indicated an indirect relationship between alexithymia and emotional eating (EES) via emotion dysregulation, and subsequently a positive conditional indirect effect whereby greater emotion dysregulation and greater self-compassion interacted, leading to greater emotional eating (EES). It was concluded that neither emotion dysregulation nor self-compassion would be appropriate targets for emotional eating interventions. The construct of ‘feeling fat’ was introduced, considered to be a proxy description used when individuals are otherwise unable to identify/describe their negative feelings, and associated with unfavourable outcomes. Existing literature is largely situated within clinical contexts, despite presence within general populations, offering an opportunity to design a brief intervention to test whether encouraging identification and description of feelings would lead to reduced state sensations of feeling fat, within the general population. The findings of the study were unexpected, as despite no significant difference in change scores across groups, the control condition elicited the greatest mean reduction in feeling fat compared to the intervention conditions. A gap in the literature examining the relationship between self-compassion and feeling fat was also examined in this final study, providing preliminary support for an inverse relationship between the traits
Encountering the Plague: Humanities Takes on the Pandemic
No full textThis edited collection features fourteen newly commissioned articles, each of which responds to the theme of plague from different disciplinary perspectives. Contributors focus on the effects of COVID-19 on everyday life, drawing also on insights from different historical experiences of plague as a way of exploring human responses to epidemics, past and present.
Each chapter opens with a different illustration that serves as a source for subsequent discussion, enabling readers to make connections between everyday objects, experiences, and broader critical debates about plague and its impact on humanity. Thought-provoking commentaries stem from a variety of humanities disciplines including archaeology, electronic literature, history, linguistics, media and cultural studies, and musicology.
Encountering the Plague explores ways in which humanities research can play a meaningful role in key social and political debates, and provides compelling examples of how the past can inform our understanding of the present
Efficient Textual Similarity using Semantic MinHashing
No full textQuantifying the likeness between words, sentences, paragraphs, and documents plays a crucial role in various applications of natural language processing (NLP). As Bert, Elmo, and Roberta exemplified, contemporary methodologies leverage neural networks to generate embeddings, necessitating substantial data and training time for cutting-edge performance. Alternatively, semantic similarity metrics are based on knowledge bases like WordNet, using approaches such as the shortest path between words. MinHashing, a nimble technique, quickly approximates Jaccard similarity scores for document pairs. In this study, we propose employing MinHashing to gauge semantic scores by enhancing original documents with information from semantic networks, incorporating relationships such as syn-onyms, antonyms, hyponyms, and hypernyms. This augmentation improves lexical similarity based on semantic insights. The MinHash algorithm calculates compact signatures for extended vectors, mitigating dimensionality concerns. The similarity of these signatures reflects the semantic score between the documents. Our method achieves approximately 64 % accuracy in the MRPC and SICK data sets
Heel Pressure Ulcers: Contributory factors and their impact on quality of life and role at the end of life in the adult population
No full textPressure ulcer prevention remains a persistent challenge in healthcare, affecting patients across all age groups, from neonates to the elderly. Pressure ulcers occur as damage to the skin and/or underlying tissue over bony prominence areas because of sustained pressure and/or shear. Heel pressure ulcers (HPUs) are the second most common type of pressure ulcer, following those at the sacrum, in most healthcare settings. There is limited evidence to support the prevention of heel pressure ulcers in the adult population. Furthermore, emerging evidence suggests differences in risk factors associated with pressure ulcers based on the specific anatomical site affected. This study therefore aims to identify and quantify the relationship between risk factors and HPUs presence and to investigate their impact on health-related quality of life (HRQoL) and prognostic significance at the end of life (EoL).
This research employed an exploratory single-centre observational study design, grounded in a pragmatic philosophy and utilising three distinct methodologies, as outlined below.
Phase 1: Risk factors for developing heel pressure ulcers in the adult population: a systematic literature review.
Phase 2: Factors associated with the presence of heel pressure ulcers in the adult population: a matched case-control study.
Phase 3: the impact of heel pressure ulcers on HRQoL and their prognostic
significance in EoL.
This thesis is the first to conduct a systematic literature review on risk factors associated with HPUs in the adult population. The review identified a significant lack of evidence, with 76.9% of the studies rated as moderate to poor quality, highlighting the need for further research. A total of 103 participants took part in Phases 2 and 3, compromising 53 patients with heel pressure ulcer(s) and 50 without. A conceptual framework for risk factors is proposed to aid in identifying those at risk and support the timely implementation of preventive strategies. HPUs negatively impact the quality of life and significantly increase healthcare resource use in acute settings. Applying evidence-based risk assessment tools (RATs) can help improve patient outcomes.
Individuals of ethnic minority backgrounds and those lacking capacity remain underrepresented in HPU research, which denies them evidence-based care and further perpetuates health inequalities. Engaging patients and the public can be essential for ensuring research is inclusive and reflects the needs of those it aims to serve. Future research needs to prioritise ethnic minority backgrounds and those lacking capacity to make HPU risk assessment and prevention processes relevant to their care
Exploring the experiences of people with obesity and post-bariatric surgery patients after three months using the mindful eating reflective practice: An interpretative phenomenological analysis
No full textBackground: Experiential dimensions of Mindful Eating Practices are scarce in the literature. Aim: The study focuses on thirteen individuals with clinical obesity and nine post-bariatric surgery patients who engaged in MERP over three months. Methods: The present research utilized Interpretative Phenomenological Analysis (IPA) as the analytical framework of interviews. Results: Four overarching themes emerged from the analysis: 1. “Enhanced Awareness of Eating": This theme underscores MERP's central emphasis on cultivating heightened mindfulness during food consumption, highlighting the importance of being present at the moment while eating; 2. “Facilitating the Transition to Healthier Eating Habits": This theme explores how MERP influences participants’ dietary choices, eating pace, portion control, and overall enjoyment of meals. It reveals that MERP encourages individuals to reflect on their eating habits and transition towards healthier choices; 3. “Diverse Perspectives on Satisfaction with MERP": Within the context of MERP, participants held varied interpretations of satisfaction. Some encountered practical limitations or engaged in reflective self-examination, while others found sensory satisfaction, enhancing their overall eating experiences; and 4. “Utilization and Development of MERP": This theme delves into participants’ patterns of using MERP. It reveals a tendency to avoid MERP in the morning, a gradual decline in its usage over time, and a preference for an electronic version of the practice. Conclusion: The MERP shows promise in improving overall eating habits by enhancing enjoyment of food, increasing awareness of body cues, promoting healthier choices, and encouraging mindful eating practices. These findings provide valuable insights for future research and the refinement of clinical tools aimed at effective weight management and the promotion of sustainable healthy eating practices by effectively addressing a significant gap in our understanding of the experiential facets of eating practices
Talent management strategies and practices in the age of digital transformation: Building bridges to successful hospitality organisations in the UK, Greece, and Hong Kong
No full textThe paper examines Talent Management strategies and practices in the age of digital transformation in 3, 4, and 5-star hotels in the UK, Greece, and Hong Kong. Using a mixed-methods research approach, the study first conducted an online survey with 63 participants from hotels in the countries mentioned above. Secondly, 20 semi-structured interviews were conducted with UK hotel professionals. The results show that hotels that are more mature in digitalisation and digitisation have created a culture of confidence and trust in technology which is conducive to integrating emerging technologies and AI in Talent Management strategies and practices. This is a pilot study that aims to understand the stage in which hotel companies are currently in the digital transformation of TM
A tale of two times: an exploration of healthcare utilization patterns before and during COVID-19 in Iran
No full textBackground
The COVID-19 pandemic has impacted global healthcare utilization patterns. This study aimed to examine the impact of COVID-19 pandemic on utilization rate of healthcare services in Iran.
Method
In this quasi-experimental study, data on the utilization rates of laboratory services, sonography exams, CT scans, MRIs, and EKGs was collected from the Social Security Organization (SSO)’s insurance information system. This data, covering 47 months prior to the pandemic and 25 months during it, was analyzed to assess the pandemic’s impact on healthcare utilization among insured individuals in Iran. The data was categorized into direct, indirect, and total sectors, and an Interrupted Time Series Analysis (ITSA) model was employed for data analysis, examining both total and sector-specific utilization rates.
Findings
The study for single group indicated that in the total sector, Utilization rate per 1000 insured significantly decreased by 25.25 for laboratory services, decreased by 3.99 for sonography, decreased by 1.08 for MRIs and decreased by 1.01 for EKGs, but increased by 2.28 for CT scans in the first pandemic month. Over following months, monthly utilization trends per 1000 insured increased significantly- laboratory services + 1.08, sonography + 0.11, CT scans + 0.12, MRIs + 0.06, and EKGs + 0.05. Pre-pandemic, monthly utilization per 1000 insured was 62.68 labs, 14.47 sonography, 0.72 CT scans, 2.06 MRIs, with all significantly higher in the indirect over direct sector except EKGs which were 2.08 higher in the direct sector. In the first pandemic month, there were significant between-sector differences per 1000 of -4.4 for sonography, + 1.89 CT scans, -1.01 MRIs and + 1.29 EKGs.
Conclusion
The COVID-19 pandemic led to a significant decline in healthcare service utilization, particularly in total and direct sectors, while CT scans remained unaffected. To address these challenges and meet patient needs, Iran’s health system should adopt alternative delivery methods like telemedicine
Role of culture and religious beliefs on non-medical help-seeking behavior among patients with chronic mental illnesses (CMIs) in Türkiye
No full textBackground:
Cultural beliefs significantly shape societal attitudes toward mental illness, and these social attitudes profoundly impact help-seeking behaviors. Therefore, it is important to focus on understanding and addressing these social behaviors.
Aim:
This study aimed to evaluate the effect of chronic mental illness interpretations based on culture and religious beliefs on non-medical help-seeking behaviors among patients in Türkiye.
Methods:
The study was conducted from September to October 2023 using an inductive qualitative approach. In-depth face-to-face interviews were carried out with individuals diagnosed with chronic mental illness and their relatives, registered in a state-owned Community Mental Health Center (CMHC) in Türkiye. Using purposive sampling, 13 individuals who met the criteria were interviewed. Thematic analysis was used to identify themes.
Results:
Three main themes and eight sub-themes were identified, including the reasons for seeking non-medical help (psychological challenges, subjective norms, physical requirements), factors contributing to seeking non-medical help (predisposing factors, enabling factors, and myths), and reflections on the benefits of non-medical practices (perceived physical benefits, perceived psychological benefits).
Conclusions:
It was concluded that individuals with chronic mental illness and their relatives living in the Eastern Anatolia Region of Türkiye engaged in non-medical help-seeking behaviors and mostly turned to traditional religious practices. Culture and religious beliefs emerged as primary factors leading patients to seek non-medical treatment approaches. Consequently, there is a perceived need to explore non-medical alternative methods across various mental health settings and with diverse samples in future research endeavors
Meta-knowledge guided Bayesian optimization framework for robust crop yield estimation
No full textAccurate pre-harvest crop yield estimation is vital for agricultural sustainability and economic stability. The existing yield estimating models exhibit deficiencies in insufficient examination of hyperparameters, lack of robustness, restricted transferability of meta-models, and uncertain generalizability when applied to agricultural data. This study presents a novel meta-knowledge-guided framework that leverages three diverse agricultural datasets and explores meta-knowledge transfer in frequent hyperparameter optimization scenarios. The framework’s approach involves base tasks using LightGBM and Bayesian Optimization, which automates hyperparameter optimization by eliminating the need for manual adjustments. Conducted rigorous experiments to analyze the meta-knowledge transformation of RGPE, SGPR, and TransBO algorithms, achieving impressive R2 values (0.8415, 0.9865, 0.9708) using rgpe_prf meta-knowledge transfer on diverse datasets. Furthermore, the framework yielded excellent results for mean squared error (MSE), mean absolute error (MAE), scaled MSE, and scaled MAE. These results emphasize the method’s significance, offering valuable insights for crop yield estimation, benefiting farmers and the agricultural sector