Saudi Cultural Missions Theses & Dissertations
Permanent URI for this communityhttps://drepo.sdl.edu.sa/handle/20.500.14154/10
Browse
154 results
Search Results
Item Restricted Engineering and Classifying Indirect Prompt Injections: A Socio-Technical Approach to AI Agent Security(Saudi Digital Library, 2026) Aldurayhim, Abdulaziz; Abudayeh, MohammadAs autonomous AI agents become more prevalent in enterprise settings, they are now susceptible to "Indirect Prompt Injection (IPI)" attacks, which embed adversarial payloads in external data sources to manipulate the agent's contextual execution and force unauthorised actions. This dissertation holds that IPI attacks constitute a socio-technical phenomenon because they exploit semantic vulnerabilities in agent architectures and psychological manipulation patterns in human communication. To address this threat, the study proposes a three-dimensional taxonomy based on payload type, injection vector, and social engineering trigger. The initial corpus comprised 1,119 instances; following rigorous preprocessing and deduplication, a refined dataset of 858 validated samples was established for modelling. A separate 1,032-sample verified dataset pack was used only as an independent reference for dataset verification, not as the definitive modelling dataset. A DeBERTa-v3-base classifier is fine-tuned to classify social engineering triggers, using RoBERTa-base as a baseline. On the OSINT subset, DeBERTa's macro-F1 score was 0.983, and the accuracy was 91.7%. The study also presents a probabilistic Confused Deputy threat model and an ACE defence architecture consisting of architectural controls, runtime detection, privilege minimisation, human-in-the-loop oversight, and practitioner education. The findings conclusively demonstrate that integrating advanced NLP detection mechanisms with socio-technical security frameworks significantly enhances enterprise resilience against IPI threats.10 0Item Restricted DESIGNING A SMART MEDICAL SUPPLY CHAIN USING MACHINE LEARNING FOR DEMAND FORECASTING IN SAUDI ARABIA(Saudi Digital Library, 2026) Alzahrani, Faisal Ahmed; Alhamad, salahThis study develops a smart medical supply chain framework using machine learning–based demand forecasting to improve inventory planning and operational efficiency in Saudi Arabia’s healthcare sector. Traditional healthcare supply chains rely on historical averages and manual planning approaches, which often result in stockouts of critical medical supplies, excess inventory, product expiration, and increased emergency procurement costs. These challenges are intensified by demand variability driven by patient admissions, seasonal trends, and large-scale events such as Hajj and public health emergencies. To address these limitations, the research adopts a Design Science Research approach and evaluates multiple forecasting models using a dataset covering ten hospitals, fifty medical supply items, and five years of demand data. Among the tested models, the Long Short-Term Memory (LSTM) neural network achieved the highest accuracy. The forecasting results were then integrated into an AnyLogic model to assess the impact on supply chain performance. The findings demonstrate significant improvements, including reduced stockout rates, lower inventory levels, decreased wastage, improved service levels, and reduced operational costs. Scenario testing under high-demand conditions confirmed the system’s ability to maintain performance and enhance resilience. Overall, the study highlights the potential of machine learning to transform healthcare logistics by enabling more accurate forecasting, better resource allocation, and improved supply chain responsiveness.4 0Item Restricted Using surveys and on-bus sensors to understand and predict bus passenger satisfaction(Saudi Digital Library, 2026) Ghulman, Mayyar; Enoch, Marcus; Kyriakopoulos, Kostas; Beechey, MattBackground: Rapid urbanisation and population growth have increased the demand for mobility in cities, leading to congestion and environmental issues such as air and noise pollution. Public transportation, particularly buses, is vital for accessibility, economic participation, and reducing social isolation, especially in areas with limited rail services. Despite these benefits, bus usage in many developed countries has stagnated at best and, in some cases, declined, while car travel is on the rise, underscoring the need to enhance bus services, make them more attractive, and increase patronage. A key challenge is the reliance on passenger perception surveys to assess bus service quality. Objective measures, such as temperature and crowding, are often collected in a fragmented manner, creating a gap in the evidence available to operators and policymakers. Without integrating perceived service quality (PSQ) with delivered service quality (DSQ), identifying practical service improvements to boost passenger satisfaction becomes obscure. Purpose: This thesis aims to investigate the impact of perceived service quality and actual delivered service quality—measured through real-time bus performance indicators— on bus passenger satisfaction and loyalty. Additionally, it evaluates predictive models that leverage real-time operational data from onboard bus sensors to predict passenger satisfaction. Methods: A quantitative research design was adopted that combined passenger perceptions with measured bus performance. Data were collected from two bus services in the UK (East Midlands): Service 90 (Nottingham–Newark) and Novus Direct (Leicester–New Lubbesthorpe). Perceived service quality was captured through an in-person, web-based customer satisfaction survey, which yielded 208 valid responses, while delivered service quality was measured using an onboard sensor system that recorded five operational indicators: punctuality, busload, speed, vibration, and temperature. The data analysis followed three stages: (1) principal component analysis reduced the perceived service quality indicators to latent dimensions; (2) the hypotheses were tested using PLS-SEM, together with multi-group analysis (MGA) to compare passenger groups; and (3) supervised machine learning was used to predict satisfaction from the operational indicators, with performance evaluated using standard classification metrics. Findings: Perceived service quality dimensions, namely pre-trip experience, on-bus environment, driver-related factors, and safety, were found to significantly contribute to predicting satisfaction, with the onboard environment being the most significant factor. Bus crowding and comfort: driving style, steering, and driver behaviour were among the most important attributes, with the highest loadings. There were no significant differences in these predictions between captive and choice riders, or between work and non-work trip groups, based on MGA. However, separate analyses of each group’s sample produced very few differences. On the other hand, in terms of delivered service quality, only actual bus load and actual speed were found to have a significant, albeit small, negative impact on satisfaction. The Random Forest classifier achieved the strongest predictive performance among the supervised classification models, with temperature, bus load, and punctuality identified as the most important predictors of passenger satisfaction. Contribution: This thesis presents a comprehensive investigation that significantly contributes to the field of public transport satisfaction through data collection and analysis. It is one of the few studies to utilise both subjective survey-based data and objective sensor-derived bus operational data to assess and predict passenger satisfaction with bus travel. Methodologically, the thesis demonstrates a practical approach for capturing, synchronising, and merging high-frequency onboard sensor data with passenger trip-level survey responses. While the integration of subjective and objective measures is established in other domains, this thesis is among the few to apply this approach to bus passenger satisfaction, using a purpose-built onboard sensor system to capture service quality at the individual-trip level and link it directly to passengers’ survey responses. This dual approach generates insights into both satisfaction and loyalty, as well as actionable operational predictions. Conclusion: This study provides bus operators and transport authorities with evidence-based methods for real-time service monitoring and targeted interventions to improve passenger satisfaction and loyalty. This research contributes to the existing body of knowledge, not only regarding buses but also for the public transport sector as a whole.5 0Item Restricted An Optimised Hybrid Framework for Predicting Learner Dropout in Massive Open Online Courses Using Machine Learning and Behavioural Analytics(Saudi Digital Library, 2026) Alghamdi, Saad Hamdan A; Ben, SohThis research addresses the persistent challenge of high dropout rates in Massive Open Online Courses (MOOCs) by developing and evaluating an enhanced predictive model capable of identifying learners at risk of disengagement. The study introduces the Integrated Stacked Ensemble Learning Dropout Prediction (ISELDP) model, which combines machine learning and meta-heuristic optimisation to improve predictive performance and interpretability. Employing an explanatory, sequential quantitative methodology, the research was conducted in two phases: (1) developing and testing the proposed model using a secondary in-session dataset, and (2) validating the model with primary behavioural and demographic data collected from 466 MOOC participants through a structured survey. The proposed ISELDP integrates multiple base learners (random forest, AdaBoost, gradient boosting, and XGBoost) with a multilayer perceptron (MLP) meta-learner, supported by an optimised feature selection framework based on a hybrid genetic algorithm–correlation feature selection (GA–CFS) technique. The evaluation across balanced and unbalanced datasets demonstrates that the enhanced ISELDP significantly outperforms benchmark models, achieving an average accuracy of 91% and an F1-score of 88%. The hybrid optimisation process effectively reduces feature redundancy, mitigates overfitting, and enhances the model’s generalisability across heterogeneous data. Complementary statistical analyses of the survey data identified motivation and engagement as the most influential predictors of MOOC retention, mediated by self-management and social influence factors, and indirectly affected by course design. These findings emphasise the centrality of motivational and behavioural dimensions in sustaining learner persistence, offering a broader understanding of dropout triggers beyond in-session activity metrics. The research contributes theoretically by defining the specifications of effective dropout prediction models that integrate deep learning and meta-heuristic optimisation. Methodologically, it advances predictive learning analytics through a validated hybrid framework that balances accuracy, interpretability, and adaptability in dynamic learning environments. Practically, it provides a data-driven foundation for designing proactive retention strategies and personalised interventions in MOOCs. Despite limitations related to data scope, computational complexity, and temporal coverage, the findings establish a replicable and extensible model for future cross-platform and longitudinal studies in predictive educational analytics. Compared to state-of-the-art approaches, this study proposes a hybrid GA–CFS feature selection method embedded in a stacked ensemble framework. The proposed approach yields a more compact feature set while enhancing prediction accuracy and robustness compared to conventional single-method techniques.21 0Item Restricted Quantifying the Human Impact on Vegetation Health in Rawdat Al Khafs, Saudi Arabia, Using Multi-Source Geospatial Data and Machine Learning(Saudi Digital Library, 2026) Alajmi, Abdulhadi; Dewan, AshrafArid and hyper-arid environments are vulnerable to land degradation and have become focal targets for large-scale afforestation initiatives combating desertification. The One Million Tree Initiative at Rawdat Al Khafs, a hyper-arid rawdah ~100 km northeast of Riyadh receiving < 50 mm annual rainfall, operates under the Saudi Green Initiative and Vision 2030. Evaluating afforestation effectiveness in such settings remains methodologically challenging, as existing frameworks rarely separate climatedriven from human-induced vegetation change in hyper-arid conditions. This study addresses these gaps through three inter-connected objectives: (i) establishing a climate-driven vegetation baseline using a Random Forest (RF) regression model trained on pre-plantation Sentinel-2 imagery and multi-source climate data (20182021); (ii) detecting and quantifying residual vegetation trends (2018–2025) that exceed this baseline and may represent a non-climatic signal conditionally attributable to afforestation; and (iii) characterising the spatial fingerprint of the detected anomaly through phase-based proximity analysis and distance-decay modelling. The RF–RESTREND framework combines 10 m Sentinel-2 imagery, CHIRPS precipitation, and ERA5-Land temperature data with Mann–Kendall trend testing, Sen's slope, Difference-in-Differences (DiD) correction, and SHAP analysis, applied to both NDVI and EVI. A persistent positive vegetation anomaly was observed within the plantation boundary from mid-2022 onwards, consistent across both indices and not fully explained by the climate counterfactual. Within the study area, 77.6% of NDVI pixels and 81.7% of EVI pixels exhibited statistically significant greening (Mann–Kendall Z > +1.96), with mean Z-scores of +3.97 and +4.45, respectively. The EVI distance-decay model (R² = 0.53; decay length = 65 m) indicated a localised boundary effect, while SHAP analysis identified spatial position and long-term climatic trends as dominant predictors, consistent with micro-topographic moistureconcentration mechanisms in rawdah ecosystems. Attribution is framed as conditional, reflecting the short pre-plantation training window and the widening climate variability of the post-plantation period. These findings provide an early satellitebased, climate-corrected assessment of an ecological response associated with the One Million Tree Initiative. The RF–RESTREND framework is reproducible and scalable to comparable arid restoration contexts, supporting Vision 2030 monitoring.17 0Item Restricted Evaluating Hybrid AI Approaches in email Spam Detection: A Literature Review(Saudi Digital Library, 2026) Alshamrani, Husam A; Sabar, NasserSpam is still one of the most serious cybersecurity issues of the day, and serves as a method of delivering phishing and malware. In this thesis, two complementary parts are integrated: First, a PRISMA-based systematic review of 55 studies that were published from 2023 to 2025 and analyze the classical, AI-based and ensemble spam detection techniques; and second, a controlled proof-of-concept experiment on the Enron-Spam corpus, where three distinct classical classifiers (Naïve Bayes with Bag-of-Words, Logistic Regression with TF-IDF, Linear SVM with TF-IDF) are compared against a stacking ensemble of the three classifiers with a Logistic Regression meta-classifier. All four configurations were very close together and obtained an accuracy of between 0.989 and 0.993, with the stacking ensemble having the highest F1-score (0.9923) and the lowest false-positive rate (0.0088); but there were no significant differences between the four configurations when a paired statistical test was not used. The thesis contribution lies in the synthesis of the various perspectives of the 2023–2025 hybrid and ensemble approaches and an internally consistent baseline comparison on a standard corpus. The study is restricted to English-language email and to lexical features; further extensions of the study using deep-learning, multilingual, and adversarial techniques are suggested.17 0Item Restricted DEEP REINFORCEMENT LEARNING BASED SEMI AUTONOMOUS CYBERATTACK DETECTION FOR INTERNET OF THINGS (IOT) SYSTEMS(Saudi Digital Library, 2026) AlRuwaili, Ibtihaj; AlQawasmi, KhaledAbstract The Internet of Things (IoT) has made today's networked world more complex and vulnerable, thus making it an appealing target for advanced cyberattacks. Limited adaptability, poor generalization and responsiveness to new and evolving attack behaviors are common problems with the traditional cybersecurity approaches that rely on static rules and supervised Machine Learning (ML). Furthermore, traditional IPSs are not equipped with the ability to adapt and learn over time in response to changing threats and vulnerabilities in the cyber landscape. To overcome the restrictions a novel Deep Reinforcement Learning (DRL) based cybersecurity framework is proposed for adaptive intrusion detection in heterogeneous enterprise-network and IoT based environment.The framework turns intrusion detection into a sequence of decisions, where a DRL agent keeps interacting with the environment and optimizes intrusion detection policies using rewards-driven learning. In order to test the diversity of the experiments and analyze the adaptive learning behavior, the framework was experimentally evaluated on two cybersecurity benchmark datasets: CICIDS2017 and TON_IoT.It was evaluated experimentally using CICIDS2017 and TON_IoT datasets, which are both widely used in the cybersecurity research community. Three DRL algorithms DDQN,PPO,DDPG were implemented and tested with various performance measures such as Accuracy, Precision, Recall, F1-Score, False positive rate (FPR) and ROC-AUC.Results of the experiments showed that the different evaluated DRL algorithms performed well in both cybersecurity test environments. However DDQN was found to be the best in terms of classification consistency, convergence stability and adaptive learning behavior in comparison to PPO and DDPG models and it was the most consistently and stably performing model. In particular, on the CICIDS2017 dataset, DDQN scored 97.3% on the Accuracy metric, 97.4% on the F1-Score metric, and 0.989 on the ROC-AUC metric, and on the TON_IoT dataset, the metrics were 95.6% for Accuracy, 95.6% for f1-score, and 0.978 for ROC-AUC. Moreover, DDQN achieved an FP rate of merely 2.1% and 3.4% in CICIDS2017 and TON_IoT respectively, suggesting a strong intrusion-classification capability under heterogeneous traffic conditions.The findings further showed that IoT environments bring in more complexity in cybersecurity with heterogeneous telemetry and dynamic communication patterns. Moreover, comparison showed that DRL-based method offers better adaptability, ongoing policy optimization, and better decision-support capability in dynamic cybersecurity scenarios.8 0Item Restricted AI-Enabled Autonomous Knowledge Extraction from Large-Scale Textual Data(Saudi Digital Library, 2026) Alharbi, Abdulrahman; Obradovic, ZoranThe rapid growth of large-scale textual data across social media platforms, news media, and scientific repositories presents both unprecedented opportunities and significant challenges for extracting meaningful insights. During global events such as the COVID-19 pandemic, understanding public discourse requires analyzing vast amounts of noisy, heterogeneous, dynamic, and geographically distributed data. At the same time, the exponential increase in scientific publications has made traditional evidence synthesis methods increasingly labor-intensive, time-consuming, and difficult to scale. Existing approaches to textual knowledge extraction often operate in isolation, lack interpretability, fail to integrate heterogeneous data sources, and do not support scalable end-to-end automation. This dissertation addresses these limitations by proposing a unified framework for AI-enabled autonomous knowledge extraction from large-scale textual data. The research introduces a comprehensive pipeline that integrate sentiment analysis, topic modeling, semantic interpretation, spatiotemporal reasoning, and multi-agent automation for scalable, robust and reproducible text analysis across heterogeneous domains. First, this work introduces TriLex, a novel unsupervised sentiment analysis framework that combines multiple lexicon-based sentiment analysis methods through weighted aggregation, majority voting, and dynamic thresholding technique to improve robustness and accuracy for short and noisy textual data. Building on this foundation, a hierarchical spatiotemporal framework is developed to capture the evolution of public sentiment across global, national, and regional scales. The framework integrates over 7 million social media posts and thousands of news articles to analyze COVID-19 vaccine discourse across time, geographic regions, and platforms. To enhance topic interpretability, this research integrates BERTopic with large language models (LLMs), enabling automated generation of coherent and context-aware topic representations for large-scale textual discourse. A cross-platform analytical framework is further introduced to examine temporal relationships between social media and news media discourse, demonstrating a bidirectional relationship in which news coverage and public discourse influence each other over time. Extending beyond discourse analysis, this dissertation introduces an Agentic AI framework that automates the end-to-end process of large-scale multilingual knowledge extraction and evidence synthesis. The proposed multi-agent system coordinates specialized agents for query generation, multilingual retrieval, metadata harmonization, title and abstract screening, and full-text analysis. Evaluated on a multilingual corpus of over 52,000 scientific records, the framework achieves high screening performance while substantially reducing processing time from months to hours, demonstrating significant improvements in scalability, robustness, and reproducibility. Collectively, this dissertation bridges the gap between analytical understanding and autonomous knowledge extraction from large-scale textual data. By integrating robust sentiment analysis, interpretable topic modeling, spatiotemporal discourse analysis, and autonomous AI systems within a unified framework, this work establishes a scalable and extensible paradigm for transforming heterogeneous textual data into actionable knowledge. The proposed methodologies are validated using real-world datasets spanning social media, news media, and scientific literature across diverse application domains.13 0Item Restricted Harmonic Impact Assessment and Prediction in a Power System with Nonlinear Loads(Saudi Digital Library, 2025) Almogargesh, Yousef Yaqoob Y; Swamidoss, SathiakumarThis thesis presents a simulation-based and machine learning framework for predicting power quality (PQ) indices in residential power systems with nonlinear loads. Four MATLAB/Simulink models representing nonlinear loads, induction motor loads, thyristor converters, and photovoltaic grid-connected systems were developed to generate voltage and current waveforms under different operating conditions. Time-domain and frequency-domain features were extracted from the simulated signals and used to train artificial neural network (ANN) models for predicting six key PQ indices: Total Harmonic Distortion (THD), Crest Factor, Form Factor, Power Factor, RMS Voltage, and Peak Voltage. Both shallow and deep neural network architectures were implemented and evaluated using statistical performance metrics including Mean Squared Error (MSE), Mean Absolute Error (MAE), and Pearson correlation coefficient. The results demonstrate that the proposed approach accurately predicts future power quality conditions while reducing the need for repeated analytical calculations. This framework provides an efficient tool for power quality monitoring and supports the development of intelligent and proactive power system management strategies.21 0Item Restricted Personalized Course Recommendations Leveraging Machine & Transfer Learning Toward Improved Student Outcomes(Saudi Digital Library, 2026) Algarni, Shrooq; Frederick, SheldonAt matriculation, university advising typically operates under tight informational constraints, often with no access to post-enrolment interaction history. We propose a unified, leakage-controlled pipeline that (i) predicts early dropout risk and (ii) generates cold-start programme recommendations using only pre-enrolment signals, with an optional early-warning variant that additionally incorporates first-term academic aggregates. The pipeline instantiates lightweight multimodal components: a tabular RNN, a DistilBERT encoder for short profile sentences, and a cross-attention fusion module, trained and evaluated end-to-end on a public benchmark (UCI id 697; n = 3630 students across 17 programmes). For dropout prediction, fusing text with numeric features yields the strongest thresh olded performance (Hybrid RNN–DistilBERT: F1 ≈ 0.9161, MCC ≈ 0.7750), while simple ensembling modestly improves threshold-free discrimination (AUROC up to ≈ 0.9488 via Stacking Ensemble, compared to ≈ 0.9459 for Weighted Ensemble). A text-only branch performs substantially worse, indicating that numeric demographics and early curricular aggregates carry most of the predictive signal at this horizon. For programme recommendation, pre-enrolment demographics alone support actionable rankings (De mographic MLP: NDCG@10 ≈ 0.5793, Top-10 ≈ 0.9380), outperforming a popularity prior by roughly 25–27 percentage points in NDCG@10; adding text yields only marginal improvements in hit rate and does not improve NDCG on this cohort. Methodologically, we apply leakage guards, deterministic preprocessing, stratified splits, and comprehensive metric reporting to enable reproducibility on non-proprietary data. Practically, the pipeline supports orientation-time triage via high-recall early warning and shortlist generation for programme selection. Overall, the results cast matriculation-time advising as a joint prediction–recommendation problem solvable with carefully engineered pre-enrolment views and lightweight multimodal models, without relying on historical interactions.13 0