APAkampurira Paul

Akampurira Paul

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Akampurira Paul has a total of 29 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Open Access

    A Deep Learning-Based User Behavior Analytics Model for Proactive Cyber Threat Detection and Risk Management: A Review

    by Paul Akampurira, Enerst Edozie, Bashir Olaniyi Sadiq, Mohammed Dahiru Buhari

    As cyber threats evolve, traditional rule-based systems struggle to detect complex and dynamic malicious behaviors, and the vulnerability of organizations to insider threats has drastically increased. Individuals entrusted with access or knowledge of the organization have become a significant concern. This review explores deep learning-based User Behavior Analytics and the integration of context-aware deep learning models for insider threat detection. Also, we investigate the use of hybrid models, including attention-based Long Short-Term Memory (LSTM) networks, which combine sequential modeling with attention mechanisms to enhance context-awareness and improve threat detection and proactive risk management. Furthermore, this paper highlights proactive risk management and dynamic interventions grounded on personalized risk profiling and user micro-segmentation. We further explore studies that give a deeper understanding of how these deep machine learning models align with ISO/IEC 27001:2022 standards, and how they can be integrated into existing frameworks to bolster proactive risk management efforts. By delivering insights into the future of AI-driven cybersecurity, this paper highlights the need to adapt to evolving threats and bolster the resilience of digital infrastructures through intelligent and adaptive security solutions.


  • Open Access

    Interpretable Discovery of Clinically Meaningful Multimorbidity Phenotypes from Routine Electronic Health Records

    by Akampurira Paul, Ibrahim Adabara, Mugisha Brian, Atuhe Aarone, Partick Aja Maduabuchi, Ibrahim Babangida Abubarkar, Maninti Venkateshwarlu, Grace Aghaji Nwankwo, Mary Agunloye, Kirungi Richard, et al.

    Multimorbidity is a growing global health challenge that complicates patient management because of heterogeneous disease combinations and complex clinical trajectories. Existing multimorbidity phenotyping approaches often rely on predefined disease groupings or conventional clustering techniques that inadequately capture latent relationships among patients, limiting the discovery of clinically meaningful phenotypes. We developed an interpretable artificial intelligence framework to discover clinically meaningful multimorbidity phenotypes from routine electronic health records by integrating structured clinical data, disease representation learning, network analysis, statistical validation, and explainable artificial intelligence. The framework was evaluated using 15,732 patient encounters and identified five distinct multimorbidity phenotypes with significantly different demographic and clinical characteristics. Statistical validation demonstrated significant differences in age (ANOVA, F = 89.27, p < 0.001), waiting time (ANOVA, F = 129.84, p < 0.001), healthcare setting (χ² = 1192.84, p < 0.001), and ICD-10 disease distributions (χ² = 777.02, p < 0.001), confirming the robustness and clinical distinctiveness of the identified phenotypes. Explainable artificial intelligence further revealed that prior healthcare utilization, laboratory investigations, medication-related variables, and disease-network embeddings were the principal determinants of phenotype membership, enabling transparent interpretation of patient stratification. By combining representation learning with disease network modelling and explainable AI, the proposed framework transforms routinely collected electronic health records into clinically interpretable multimorbidity phenotypes, providing a scalable foundation for precision medicine, clinical decision support, and population health management.

  • Open Access

    Risk-confidence co-estimation in insider threat detection using uncertainty-aware deep sequential models for proactive risk management

    by Akampurira Paul, Sadiq O. Bashir, Buhari Dahiru, Mandela Ngaira, Maninti Venkateswarlu

  • Open Access

    A Digital Competence-based Framework for Evaluating and Enhancing Electronic Health Record Utilization in Hospital Settings: Evidence from a Low- and Middle-Income Country

    by Brian Mugisha, Joe Mutebi, Maninti Venkateswarlu, Paul Akampurira

    Background Electronic Health Record (EHR) systems are critical for improving healthcare delivery; however, their effective utilization remains limited in many settings due to inadequate digital competence among healthcare professionals. Methods This study developed a digital competence–based framework using empirical data from healthcare workers in hospitals in the Greater Bushenyi region of Uganda. A mixed framework development approach was applied, integrating statistical analysis, theoretical models (Self-Determination Theory, DigComp), and expert validation using the Delphi technique. Results The findings revealed that human, organizational, technical, and system factors significantly influence digital competence and EHR utilization. Human factors emerged as the strongest predictor of healthcare service delivery outcomes. The proposed framework positions digital competence as a central mediating mechanism linking EHR utilization determinants to service delivery. Conclusions The framework provides a practical tool for assessing workforce readiness, identifying competence gaps, and guiding targeted interventions to improve EHR utilization. It is particularly relevant for healthcare systems in low- and middle-income countries.

  • Open Access

    Improving Health Workers' Competence in Electronic Health Records Utilization for Quality Service Delivery: A Systematic Review

    by Brian Mugisha, Venkateswarlu Maninti, Joe Mutebi, Paul Akampurira, Latifat Aderonke Adeyemi

    Introduction Electronic Health Records (EHRs) are important for improving healthcare quality, but many health workers do not fully use them due to limited skills. This reduces their benefits and can affect patient care. Methods A systematic literature review was conducted in accordance with PRISMA guidelines. Multiple databases, including PubMed, Scopus, JMIR, JAMIA, and BMC, were searched for empirical studies published between 2020 and 2025. Studies were selected using predefined Population–Intervention–Outcome–Study design (PIOS) criteria. A total of 34 studies met the inclusion criteria. Data were extracted using a standardized form, and methodological quality was assessed using the Joanna Briggs Institute appraisal tools. Due to heterogeneity in study designs and outcome measures, findings were synthesized using a thematic analysis approach informed by Social Cognitive Theory and Adult Learning Theory. Results Five main approaches were found to improve EHR skills: role-based training, simulation training, peer support (super-users), short on-the-job learning (microlearning), and follow-up training after system use. These methods helped improve efficiency, reduce errors, increase confidence, and lower burnout among health workers. Discussion & Conclusion EHR competence is not just about technical skills, it also involves training, support, and continuous learning. Combining different training methods works best. Healthcare organizations should invest in ongoing training to improve patient safety, service quality, and staff well-being.

  • Open Access

    A User-Behavior Micro-Segmentation Framework for Cyber security in Work-From-Home Environments

    by Atuhe Aarone Mike, Akampurira Paul, Wilson Tumuhimbise, Richard Ntwari, Aggrey Obbo

    Background The rapid adoption of work-from-home (WFH) arrangements in higher education has increased dependence on personal devices and home networks, exposing universities to cybersecurity risks influenced by user behavior, usability challenges, and environmental constraints. This study aimed to develop a User-Behavior Micro-Segmentation Framework (UBMSF) tailored to the contextual and operational realities of remote work in Ugandan universities. Methods A quantitative descriptive research design grounded in the Theory of Planned Behavior (TPB) and Design Science Research (DSR) principles was employed. Data were collected using structured questionnaires administered to academic, administrative, and information technology staff from selected Ugandan universities. Descriptive statistics, correlation analysis, multiple regression, and k-means clustering were used to analyse cybersecurity behavior, usability difficulty, digital fatigue, and behavioral patterns. Results The findings revealed that perceived behavioral control, usability difficulty, and digital fatigue were significant predictors of cybersecurity behavior. Regression analysis showed that usability difficulty and fatigue negatively affected secure practices, while clustering analysis identified distinct user segments with varying levels of risk exposure, capability, and behavioral consistency. These findings informed the development of the UBMSF, which integrates behavioral, technological, and contextual dimensions to support adaptive cybersecurity interventions. Conclusions The study demonstrates that cybersecurity behavior in remote work environments is shaped by both psychological and contextual factors. The proposed framework provides a structured approach for identifying user behavior patterns and aligning cybersecurity interventions with user capability, usability conditions, and environmental realities in higher education institutions.

  • Open Access

    A Deep Learning-Based User Behavior Analytics Model for Proactive Cyber Threat Detection and Risk Management: A Review

    by Paul Akampurira, Enerst Edozie, Bashir Olaniyi Sadiq, Mohammed Dahiru Buhari

    As cyber threats evolve, traditional rule-based systems struggle to detect complex and dynamic malicious behaviors, and the vulnerability of organizations to insider threats has drastically increased. Individuals entrusted with access or knowledge of the organization have become a significant concern. This review explores deep learning-based User Behavior Analytics and the integration of context-aware deep learning models for insider threat detection. Also, we investigate the use of hybrid models, including attention-based Long Short-Term Memory (LSTM) networks, which combine sequential modeling with attention mechanisms to enhance context-awareness and improve threat detection and proactive risk management. Furthermore, this paper highlights proactive risk management and dynamic interventions grounded on personalized risk profiling and user micro-segmentation. We further explore studies that give a deeper understanding of how these deep machine learning models align with ISO/IEC 27001:2022 standards, and how they can be integrated into existing frameworks to bolster proactive risk management efforts. By delivering insights into the future of AI-driven cybersecurity, this paper highlights the need to adapt to evolving threats and bolster the resilience of digital infrastructures through intelligent and adaptive security solutions.

  • Open Access

    A Framework for Evaluating the Usability of Mobile Learning Applications in Universities

    by Akampurira Paul

    As the use of mobile devices overwhelmingly outpace that laptops and desktop computers today, high levels of usability of these mobile devices are a major supporting factor. The trend extends deeply into teaching and learning as mobile learning (a new form of electronic learning) takes shape, introducing an eminent aspect of mobility. Usability defines the aspects of quality in use of the software products; hence usability testing is a major concern in developing organizations for the success of product implementation and use. Most existing usability evaluation models were developed for desktop software development. Consequently, currently available models are not about mobile learning in particular, which also presents a gap that we attempted to fill. The researcher in this study developed a model that estimates usability as a function of aggregated usability influencing factors. The developed model combines important factors from other available models and encompasses most that support mobile learning to come up with a more comprehensive model. To evaluate the effectiveness of the developed model, a mobile learning prototype application was developed, tested and implemented alongside a task list for an objective study and survey questionnaire for a subjective study. The feedback from the experiment and survey was then used to assess and validate the prototype application in terms of high, average or low levels of usability using an advanced statistical tool. The results could be used as guidelines to developing organizations to produce more appropriate applications for mobile learning with high levels of usability