Saudi Cultural Missions Theses & Dissertations
Permanent URI for this communityhttps://drepo.sdl.edu.sa/handle/20.500.14154/10
Browse
9 results
Search Results
Item Restricted Exploring the Security Landscape of AR/VR Applications: A Multi-Dimensional Perspective(University of Central Florida, 2025) Alghamdi, Abdulaziz; Mohaisen, DavidThe rapid evolution of Augmented Reality (AR) and Virtual Reality (VR) technologies on mobile platforms has significantly impacted the digital landscape, raising concerns about security and privacy. As these technologies integrate into everyday life, understanding their security infrastructure and privacy policies is crucial to protect user data. To address this, our first study analyzes AR/VR applications from a security and performance perspective. Recognizing the lack of benchmark datasets for security research, we compiled a dataset of 408 AR/VR applications from the Google Play Store. The dataset includes control flow graphs, strings, functions, permissions, API calls, hexdump, and metadata, providing a valuable resource for improving application security. In the second study, we use BERT to analyze the privacy policies of AR/VR applications. A comparative analysis reveals that while AR/VR apps offer more comprehensive privacy policies than free content websites, they still lag behind premium websites. Additionally, we assess 20 U.S. state privacy regulations using the Coverage Quality Metric (CQM), identifying strengths, gaps, and enforcement measures. This study highlights the importance of critical privacy practices and key terms to enhance policy effectiveness and align industry standards with evolving regulations. Finally, our third study introduces a scalable approach to malware detection using machine learning models, specifically Random Forest (RF) and Graph Neural Networks (GNN). Utilizing two datasets—one with Android apps, including AR/VR, and Executable and Linkable Format (ELF) files—this research incorporates features such as API call groups and Android-specific features. The GNN model outperforms RF, demonstrating its ability to capture complex feature relationships and significantly improve malware detection accuracy. This work contributes to enhancing AR/VR application security, improving privacy practices, and advancing malware detection techniques.27 0Item Restricted Intelligent Diabetes Screening with Advanced Analytics(University of Birmingham, 2024) Aldossary, Soha; Smith, PhillipDiabetes mellitus is a prevalent chronic disease with significant health implications worldwide. This project aimed to mitigate this pressing public health concern by using machine learning techniques and deep learning algorithms. I also established an online platform at which patients can enter their test results and health information and receive real-time diabetes detection and dietary recommendations based on their health profiles. Research has illustrated that models such as Gradient Boosting, Random Forest and Decision Trees perform well in diabetes prediction due to their ability to capture complex nonlinear relationships and handle diverse input features. Therefore, this project incorporated these models with others, such as the Support Vector Classifier and AdaBoost. Additionally, deep learning models, including Neural Networks, were utilised to explore intricate relationships within diabetes-related indicators. Notably, the Gradient Boosting model achieved an impressive accuracy of 99%, with 99% precision, 97% recall and 97% F1-score. To implement these solutions, I used Python as the programming language, employing libraries such as scikit-learn, NumPy, Pandas and Matplotlib, while Streamlit served as the app’s framework.18 0Item Restricted EAVESDROPPING-DRIVEN PROFILING ATTACKS ON ENCRYPTED WIFI NETWORKS: UNVEILING VULNERABILITIES IN IOT DEVICE SECURITY(University of Central Florida, 2024-08-02) Alwhbi, Ibrahim; Zou, ChangchunThis dissertation investigates the privacy implications of WiFi communication in Internet-of-Things (IoT) environments, focusing on the threat posed by out-of-network observers. Recent research has shown that in-network observers can glean information about IoT devices, user identities, and activities. However, the potential for information inference by out-of-network observers, who do not have WiFi network access, has not been thoroughly examined. The first study provides a detailed summary dataset, utilizing Random Forest for data summary classifica- tion. This study highlights the significant privacy threat to WiFi networks and IoT applications from out-of-network observers. Building on this investigation, the second study extends the research by utilizing a new set of time series monitored WiFi data frames and advanced machine learning algorithms, specifically xGboost, for Time Series classification. This extension achieved high accuracy of up to 94% in identifying IoT devices and their working status, demonstrating faster IoT device profiling while maintaining classification accuracy. Furthermore, the study underscores the ease with which out- side intruders can harm IoT devices without joining a WiFi network, launching attacks quickly and leaving no detectable footprints. Additionally, the dissertation presents a comprehensive survey of recent advancements in machine- learning-driven encrypted traffic analysis and classification. Given the challenges posed by encryp- tion for traditional packet and traffic inspection, understanding and classifying encrypted traffic are crucial. The survey provides insights into utilizing machine learning for encrypted network traffic analysis and classification, reviewing state-of-the-art techniques and methodologies. This survey serves as a valuable resource for network administrators, cybersecurity professionals, and policy enforcement entities, offering insights into current practices and future directions in encrypted traffic analysis and classification.30 0Item Restricted GIS-Based Modeling of Shallow Groundwater Potential in Arid Regions under changing Climate and Future Water Demands: a case study of Al-Madinah, Saudi Arabia.(University of York, 2024-06-14) Alharbi, Ohood; McClean, Colin; Sakai, MarcoInvestigating water resources in arid regions is essential for managing water scarcity's unique challenges in these environments. GIS and remote sensing approaches have been applied here to model and analyse three main aspects: mapping potential groundwater zones, assessing climate change impacts, and examining future water needs under socio-economic scenarios. A fuzzy-frequency ratio model and a logistic regression model successfully delineated the potential groundwater zones. An ensemble of models performed well (Best model AUC = 0.943). Soil type was the most important factor in driving both models. The spatial distribution of very high potential groundwater areas in Al-Madinah is primarily compatible with volcanic lava areas with Lithosols and Calcic Yermosols soils. Assessing climate change under IPCC RCPs scenarios (2021-2100, RCP4.5 and RCP8.5) revealed that the temperature and Reference Evapotranspiration (ET0) rate of Al-Madinah is expected to continue to increase although rainfall may also increase by around 18.74% or 22.81 mm (2081-2100, RCP8.5) compared to 1970-2018. Such an increase might not have a pronounced effect on enhancing groundwater availability due to raising temperature (2°C) and ET0 (359.70 mm) with a higher probability of drought events indicated by the Standardised Precipitation Evapotranspiration Index (SPEI). Increases with higher water accumulation opportunities are predicted at 2081-2100 (RCP8.5). However, changes in potential groundwater zones using the Topographic Wetness Index (TWI) weighted by Rainfall are expected to show a small quantitative increase with the greatest addition of suitable potential zones also estimated for 2081-2100 under RCP8.5 (logistic regression = 19296km²) Analysing water needs in Al-Madinah city under the Impact of Population, Affluence, and Technology (IPAT) model confirmed that population was the most important factor in explaining water consumption trends. Water demand is projected to increase by up to 28% under IPCC_ SSP scenarios. These findings should aid in developing water resources management strategies and sustainable decision-making.37 0Item Restricted Predicting Drugs Metabolized by Cytochrome Enzymes(University of Glasgow, 2023-12-06) Alshammari, Mariam; Lever, jakeIn the era of rapid technological evolution, embracing the strength of machine learning, deep learning, and other computational approaches merged with biological and biochemical domains has enhanced multiple medical applications. Drug discovery is one of the fields that have been rapidly developing. This project will focus on predicting drugs metabolized by Cytochrome enzymes. Therefore, using machine learning, deep learning, and pre-trained approaches would illustrate the strength of recent computational methods used in the medical field; consequently, it will reduce the limitations of traditional techniques in drug development by limiting the cost and time during clinical trials. This study will prepare the dataset to extract descriptors and build Logistic Regression, Support Vector Machine, Random Forest, Recurrent Neural Networks, ChemBERTa, and Galactica, along with parameter tunning to evaluate the best model through ROC Curve, Confusion Matrix, and F1-score. This proposed study shows that random forest outperformed other models with a 0.907 f1-score.21 0Item Restricted Measuring The Quality of Wikipedia Articles Among Different Topics(Saudi Digital Library, 2023-11-23) Aljohani, Thamer; Niesen, JitseWikipedia, a globally famous online encyclopedia, offers millions of articles across diverse topics. Its open editing policy, allowing contributions from volunteers, has made it a valuable resource. However, its reliability has been questionable, particularly in academic circles. To enhance the understanding of Wikipedia’s quality, and due to the difficulty of the assessment of quality in Wikipedia’s approach, this study presents an innovative approach to evaluate article quality. This study aims to create a quantifiable simple model based on measurable attributes, such as the length of articles, the number of references, and the number of edits. This model facilitates the calculation of article quality and subsequent assignment of quality classifications. As a result, the model proposed in this study shows an approximate accuracy equal to a random forest model which is considered a complex model. Furthermore, the research explores variations in article quality across various topics, shedding light on topics where high-quality content is prevalent and areas that require improvement. Data was collected from the Wikipedia API, and based on these measurable features, quality assessments were made. The findings indicate that Astronomy topics have a higher level of quality, while Language topics have a lower proportion of high-quality topics. These findings suggest that the attributes used to measure quality in this study are sufficient and efficient for assessing article quality on Wikipedia. Moreover, the study highlights the articles that need the experts to focus their efforts on improving articles related to topics such as Language, Business, or Mathematics to enhance the overall quality of content in these topics.10 0Item Restricted Sentiment Analysis in Online Social Networks(SDL, 2023-05-19) Assery, Ahmad Ali; Zarbaf, MonasadatThe convenience of being able to shop from home has led to the rise of e-commerce in today's highly digitized society. Before buying anything online, customers are required to read hundreds of reviews. Tracking and analyzing customer feedback may be challenging when there are millions of online reviews for a single product. However, in today’s age of machine learning, if a model were used to polarize and comprehend from them, thousands of input and information might be gained quickly and easily. As a result, sentiment analysis has emerged as a distinct field of research that integrates NLP and text analytics to identify and categorize the emotional tone of written content. In this dissertation, we investigate the difficulty of labelling reviews as positive, negative, or neutral. For massive amounts of supervised data, like those seen in the Amazon dataset, we have found success using KNN, Logistic regression, and Random Forest Classifiers. Meanwhile, the greatest results were obtained using the Logistic and random forest classifiers, and we plan to develop a web application using these models to categorize the reviews in real time. Finally, this research delves into sentiment analysis and opinion mining with regards to product feedback.39 0Item Restricted Heterogeneous Machine Learning Ensembles for Predicting Train Delays(2023-06-07) Al Ghamdi, Mostafa; Wang, WenjiaTrain delays are a serious problem in the UK and other countries. Much research has gone into developing methods for predicting train delays. Most of these methods use only single models or homogeneous ensembles and their performance in terms of accuracy and consistency in general is unsatisfactory. We have therefore developed heterogeneous ensembles that use different types of regression models with an aim of improving their prediction performance. We first looked at a wide range of base-learner models, including the state-of-the-art methods, Random Forest and XGBoost. Overall, our ensembles were more accurate than any of these single models. We developed two methods for model selection when building the ensemble, the first uses accuracy and the second uses accuracy and diversity. We found that using accuracy resulted in the most accurate ensembles. We adapted the Coincident Failure Diversity measure for regression and compared its effectiveness with other diversity measures. While it proved the best, overall, we found no relationship between ensemble accuracy and diversity in the regression context. We also investigated the effect of ensemble size. We compared the performance of our ensembles with the deep learning methods CNN and Tabnet and found that our ensembles were more accurate. However, ensembles of deep learning models proved to be more accurate than those of single machine learning models. We tested our ensembles using a different set of train delay data and found that they produced more accurate and consistent results, indicating that our methods generalise well to new data.22 0Item Restricted Investigating the Impact of Adaptive Façades on Energy Performance Using Simulation and Machine Learning(Cardiff University, 2023-04-05) Alammar, Ammar; Jabi, Wassim; Lannon, SimonBuildings consume approximately 40% of the world's primary energy, and half of this energy demand stems from space cooling and heating. To meet the targets of designing high performance buildings, intelligent solutions need to be integrated into the design process of buildings to achieve indoor environmental comfort and minimize energy consumption. In particular, the building façade plays a crucial role, as it acts as a separator element that can control the indoor environment and energy performance. This is even more important in buildings with extensive glazing systems particularly in harsh, hot climates. As stated in the literature, buildings are exposed to dynamic environmental factors that change continuously throughout the day and the year. Nonetheless, regardless of the climatic variations, building skins have been typically designed as static envelopes, which are limited in terms of their responsiveness to indoor or outdoor environmental conditions. In contrast, adaptive façades (AFs) are flexible regarding the adaptability of the system to climatic conditions enabling them to respond to short-term changes in the environment. From an environmental viewpoint, it is essential to reduce the energy consumption of buildings and mitigate their environmental impacts. Numerous innovative building envelope technologies have been developed to improve indoor comfort and reduce the environmental impact of buildings during their life cycle. As stated in the literature, AFs can make a major and practical contribution to achieving the worldwide zero-energy building targets and sustainability of our cities. In practice, assessing the performance of AFs during the early stages of the design is still a challenging task due to their time-varying dynamic behaviour. Most current building performance tools (BPS) were originally developed to assess fixed façades where changes to the geometry of the façade are not taken into consideration during simulation. To that end, adaptive systems require a more complex workflow that can correctly predict their performance. This research is intended to assist architects and façade specialists in two main aspects; firstly, an algorithmic framework was developed to predict the energy performance of AFs in the early design stages. The algorithmic workflow creates a link between plug-ins including the Ladybug and Honeybee tools, and Energy Plus for running the simulation with the built- in tool energy management system (EMS) to program a code to actuate the AF system in an hourly time step. The workflow considers the time-varying dynamic behaviour of AFs based on different environmental parameters. The aim is to accurately evaluate the potential of AFs in the energy performance of an office tower. Secondly, by exploring the complexity and limitation of current tools, a novel method is proposed to assess the energy performance of AFs using machine learning (ML) techniques. Two different ML models, namely, an artificial neural network (ANN) and a Random Forest (RF), were developed to predict the energy performance of AFs in the early design stages in a significantly faster time compared to simulation. The surrogate models were trained, tested, and validated using the generated synthetic database by simulation (hourly cooling loads of AF and hourly solar radiation). During the training phase, a hyperparameters tuning procedure was carried out to select the most suitable surrogate model. By comparing the static shading system with AFs in terms of energy consumption, the results confirmed that the AFs were more effective in terms of cooling load reduction compared to static façades where cooling loads were reduced by 34.6%. The findings also revealed that the control scenario that triggered both incident solar radiation and operative temperature in a closed loop mechanism performed better than other control scenarios. Regarding the surrogate models, this research found that ML techniques can predict the hourly cooling loads of AFs with a high level of accuracy in the range of 85% to 99%. In particular, the RF model showed a 17% improvement in R2 accuracy over the ANN model in predicting the hourly cooling loads of AFs.19 0