Effat University Institutional Repository
Not a member yet
1865 research outputs found
Sort by
Algorithms for Feature Selection (3rd Edition)
No full textAccurate photovoltaic (PV) power forecasting is essential for grid integration, particularly in maritime climates with dynamic weather patterns. This study addresses high-dimensional meteorological data challenges by systematically evaluating 32 variables across four categories (solar irradiance, temperature, atmospheric, hydrometeorological) for day-ahead PV forecasting using long short-term memory (LSTM) networks. Using six years of data from a 350 kWp solar farm in Scotland, we compare satellite-derived data and local weather station measurements. Surprisingly, downward thermal infrared flux—capturing persistent atmospheric moisture and cloud properties in maritime climates—emerged as the most influential predictor despite low correlation (1.93%). When paired with precipitation data, this two-variable combination achieved 99.81% R2, outperforming complex multi-variable models. Satellite data consistently surpassed ground measurements, with 9 of the top 10 predictors being satellite derived. Our approach reduces model complexity while improving forecasting accuracy, providing practical solutions for energy systems.
Keywords: deep learning; forecasting; long short-term memory; mean absolute; meteorological variablesEffat Universit
2025 22nd International Learning and Technology Conference (L&T)
No full textThis study presents a systematic evaluation of deep learning architectures for photovoltaic (PV) power forecasting, comparing nine model configurations across three architectures (MLP, LSTM, CNN) and three optimizers (Adam, RMSprop, Adagrad). Using six years of hourly data from a 350kWp grid-connected PV system in Scotland, we demonstrate that architectural choice and optimizer selection significantly impact forecasting accuracy. The LSTM-RMSprop configuration achieved superior performance with RMSE of 2.651 kWh and MAE of 1.197 kWh, showing a 90.29% coefficient of determination (R2). This outperforms both CNN (RMSE: 2.767-2.902 kWh) and MLP architectures (RMSE: 3.104-3.115 kWh) across all optimizers. Our main contributions include: (1) comprehensive optimization of model architectures through hyperparameter evaluation, revealing optimal configurations for each model type; (2) systematic evaluation of optimizer impact, demonstrating RMSprop's superiority for LSTM with improved accuracy across architectures; (3) detailed error analysis showing model stability across different conditions, with NRMSE ranging from 31.15% to 39.07%; and (4) practical insights into computational requirements, where CNN architectures achieve fastest training times (2.3-3.8 minutes/epoch) compared to LSTM (3.9-5.9 minutes/epoch) and MLP (36.54-46.72 minutes/epoch). Results demonstrate that LSTM architectures with appropriate optimization can outperform simpler models in PV power forecasting, providing valuable guidance for practical implementations.
Author Keywords
photovoltaic power forecasting
deep learning
hyperparameter optimization
time series prediction
optimization algorithmsEffat Universit
Assessment of obstructive sleep apnea rate and associated factors among Lebanese adults: a cross-sectional study.
No full textTo estimate the rate of obstructive sleep apnea (OSA) and associated factors among Lebanese adults.A cross-sectional study was conducted in February-April 2022; 1,202 Lebanese adults were enrolled. The sample was collected among several Lebanese governorates through an anonymous online self-reported questionnaire. The STOP-BANG questionnaire was used as a screening tool to estimate the OSA risk. "Low OSA risk" is established with 0-2 positive answers, "Intermediate OSA risk" is established by 3-4 positive answers while "High OSA risk" is established by 5-8 positive answers.This study showed that 743 (62.4%) of the sample had low risk for OSA, 357 (30.0%) had moderate, and 90 (7.6%) had high risk for OSA. Older age (aOR = 1.05) and having hypertension (aOR = 7.80) were associated with higher odds of moderate OSA. Female sex (aOR = 0.04) was significantly associated with lower odds of moderate OSA compared to males. Older age (OR = 1.17), higher BMI (OR = 1.14), hypertension (OR = 18.55), and having severe COVID-19 infection compared to mild (OR = 4.30) were significantly associated with higher odds of high OSA, whereas female sex (OR = 0.002) and being married compared to single (aOR = 0.23) were associated with lower odds of high OSA.This study showed that most Lebanese adults have low risk for OSA. It also confirmed that sex, age, obesity, hypertension, and severe COVID-19 were associated with high OSA risk. These associated factors call for future research exploring the causes including the potential effect of social, economic, and political instability, local customs, and environmental factors. Moreover, future health campaigns should be assigned to increase awareness among Lebanese population regarding the prevention of OSA through lifestyle modifications
GutGaurd: IBS Medical Chatbot
No full textGutGuard is an intelligent web-based system designed to assist individuals
with Irritable Bowel Syndrome (IBS) by providing them with personalized health
guidance through the use of artificial intelligence. The core idea behind the project is
to offer patients a reliable, easy-to-use digital assistant that helps them understand
their medical reports, plan meals based on their condition, and monitor their
symptoms over time. The system was developed using a layered architecture that
integrates frontend interfaces, backend processing, medical document analysis, and
AI-driven interaction.
The chatbot is powered by the Mistral-7B [13] language model, which is
hosted locally through the Ollama [13] framework. Users can upload their medical
reports in PDF format, and the system processes these documents using Optical
Character Recognition (OCR) to extract the text content. It then applies Medical
Entity Recognition (MER) to identify important biomarkers, such as CBC results and
metabolic indicators. The extracted information is cleaned, validated, and structured
in a way that allows the language model to understand the context and generate
relevant, medically informed response
Empowering Smart Environments through Proactive Vulnerability Detection
No full textAs the Internet of Things (IoT) continues to grow, ensuring the security of interconnected
devices becomes increasingly vital. Hythrmia is an advanced IoT security scanning tool
developed to identify vulnerabilities in smart devices across multiple communication protocols, including WiFi, Bluetooth, Z-Wave, and Zigbee. The tool scans local networks to
detect potential security risks, such as open ports, weak passwords, outdated firmware,
and known vulnerabilities (CVEs). Currently, Hythrmia focuses on WiFi and Bluetooth
devices, o↵ering real-time discovery and analysis of connected devices. The tool employs
Nmap for port scanning and Hydra for brute-force testing to identify vulnerabilities like
weak or default passwords. Additionally, it integrates the Vulners API to cross-reference
detected devices with a database of known vulnerabilities, enabling users to quickly assess
and mitigate security risks. Initial tests demonstrated show that Hythrmia’s WiFi and
Bluetooth scanning capabilities e↵ectively detected devices with weak or default passwords. The tool successfully identified several common vulnerabilities, across a range of
smart devices. For example, during a test involving several Dahua cameras, the tool was
able to exploit weak passwords and gain unauthorized access to camera streams, highlighting significant security gaps in default configurations. These findings underscore the
importance of securing IoT devices within local networks. Although the tool’s full potential includes support for Z-Wave and Zigbee, the current focus on WiFi and Bluetooth has already demonstrated Hythrmia’s significant contribution to enhancing network security.
This project highlights the increasing demand for IoT security tools and demonstrates
how Hythrmia serves as an accessible solution for users to proactively manage and secure
their smart environments. Future work will aim to expand support for additional protocols, introduce automation features for continuous scanning, and enhance the overall
e↵ectiveness of the tool as a comprehensive IoT security solutio
[PDF] from tandfonline.com Psychometric properties of an Arabic translation of the Child–Adolescent Perfectionism Scale–short form (CAPS-9)
No full textThe objective of the study was to examine the psychometric properties of the Arabic version of the nine-item Child–Adolescent Perfectionism Scale (CAPS-9). Five hundred and five adolescents participated in this cross-sectional study (mean age = 16.66; 52.7% females). The two-factor model [self-oriented perfectionism (SOP) and socially prescribed perfectionism (SPP)] indicated a good fit. Internal reliability was adequate for the SPP (ω = 0.82/α = 0.81) and SOP (ω = 0.81/α = 0.80) subscales. Convergent validity was supported by an adequate average variance extracted value. Measurement invariance across gender was supported at the configural, metric, and scalar levels, with no significant difference between males and females. Concurrent validity was demonstrated since higher SPP was significantly associated with higher distress, eustress, physical aggression, anger, and hostility. Higher SOP was significantly associated with higher eustress and anger. The present results demonstrate the usefulness of the Arabic version of the CAPS-9
A Comparative Analysis of Statistical Modeling and Machine Learning Techniques for Predicting the Lifetime of Light Emitting Diodes From Accelerated Life Testing
No full textThis study re-examines the failure behavior of high-brightness blue LEDs under accelerated life testing using multivariable stress models. Current, temperature, RH, and their interactions are analyzed, revealing that the lognormal distribution fits lifetime data better than the Weibull. The Intel model outperforms the Peck model for stress-life relationships. Machine learning is also explored for lifetime prediction, though its limitations under normal conditions are noted.This work uses multivariable life stress models to revisit the catastrophic failure of high-brightness blue light emitting diodes (LEDs) under accelerated life testing (ALT). The stress factors, current, temperature, relative humidity (RH), and their interactions are considered in lifetime studies. First, we show that the lognormal distribution fits the experimental data much better than the Weibull distribution using the standard Kolmogorov-Smirnov test. Furthermore, the best life-stress relationship is the Intel model rather than the peck model used by Nogueira et al. (2016). Additionally, based on the accelerated data, machine learning (ML) techniques are employed to predict the lifetime of LEDs under normal operating conditions. However, the study highlights the limitations of ML in accurately predicting lifetime
The Relationship Between Early Maladaptive Schemas and Intimate Partner Violence Against Women: The Moderating Effect of Childhood Trauma.
No full textThe study objectives were to test the hypothesis that childhood trauma moderates the associations between early maladaptive schemas (EMS) and intimate partner violence (IPV) among women in the specific Lebanese patriarchal context. This is a cross-sectional study that was conducted on Lebanese women between September and December 2018; 1,655 participants enrolled in this study were from all of Lebanon's governorates and were selected using an equitable representative sample. The "Disconnection and Rejection" EMS domain showed the strongest correlations with both physical and nonphysical IPV ( = .46 and = .51, respectively) in our sample. Moderation analyses findings showed that at low, moderate, and high levels of childhood trauma, greater endorsement of the "Disconnection and Rejection" schema domain was strongly linked to more severe physical and nonphysical IPV. Furthermore, childhood trauma (only at high levels) emerged as a significant moderator in the link between the "Impaired autonomy and performance" domain and physical/nonphysical IPV. High levels of childhood trauma significantly moderated the association between "Other directedness" and nonphysical IPV. Finally, low levels of childhood trauma moderated the link between "Over-vigilance and Inhibition" and physical IPV. Given that EMS are known to be resistant to change, identifying childhood trauma as a moderator in the link between certain specific EMS and IPV can provide novel avenues for the prevention IPV and its long-lasting detrimental consequences. Childhood trauma can be regarded as a target for prevention and intervention, as it can assist in mitigating correlations between EMS and IPV occurrence
