From research to decisions.
From technology to institutional impact.

This study combines physiological signals, facial images, MRI, and clinical text for automated pain detection. Modality-specific networks feed feature-refinement and particle-swarm selection and weighting stages before fusion. The reported experiments achieved 96.2% accuracy on the study datasets, supporting further investigation of multimodal assessment rather than establishing independent clinical validation.

Authors: Karim Gasmi, Sarah Almaghrabi, Ibrahim Rizqallah Alzahrani, Adil Mahmoud Mohamed Mahmoud, Alameen Abdalrahman Eltoum, Abdullah Sawdi Alenzi, Osama R Shahin, Ahmed I Taloba, Fatma Al Zahraa Hamed Abdel Hameed

Editorial summary based on the publisher abstract or original publication metadata.

The study examines why awareness of AI-enabled phishing may not consistently translate into protective action. A cross-sectional survey of 300 university students is analysed with PLS-SEM. Its socio-technical framework connects perceived awareness, institutional support, feedback, adaptive response capacity, and self-reported protective behaviour. The available abstract concerns perceived relationships, rather than a controlled test of actual phishing resistance.

Authors: Ahmed Abdallah Abaker, Ibrahim R Alzahrani, Noha Abdulkhaliq M Alharbi

Summary prepared from the abstract indexed in Google Scholar.

An integrated framework combines system dynamics, agent-based modelling, and network analysis for health-system planning under uncertainty. An Adaptive Policy Intelligence Layer supports feedback-driven monitoring, scenario evaluation, and policy adjustment. Simulations covering demand shocks, resource constraints, and digital transformation indicate improved resilience and resource deployment relative to single-method approaches within the evaluated scenarios.

Authors: Ahmed Abdallah Abaker, Khalid Aldriwish, Ibrahim Rizqallah Alzahrani, Daifallah Zaid Alotaibe

Editorial summary based on the publisher abstract or original publication metadata.

Using a cross-sectional survey of 312 respondents in Saudi and GCC organizations, the study examines sustainable computing and readiness for environmental, social, and governance reporting. PLS-SEM and multi-group analysis identify relationships involving organizational drivers and barriers. Sustainable computing mediates the conversion of fragmented digital data into structured ESG reporting, with differences across sectors.

Authors: Ahmed Abaker, Asim Seedahmed Ali Osman, Aeshah Alotaibi, Ibrahim Rizqallah Alzahrani, Daifallah Zaid Alotaibe

Editorial summary based on the publisher abstract or original publication metadata.

GeneticNAS searches automatically for neural architectures for autism spectrum disorder classification while balancing predictive performance and computational complexity. It combines a memory-efficient genetic algorithm, a varied operation search space, and adaptive fitness evaluation. The paper reports 76% lower GPU-memory consumption and 95.23% classification accuracy in its experiments, motivating further work in resource-constrained settings.

Authors: Abdullah R Alzahrani, Dabiah Alboaneen, Ibrahim R Alzahrani

Editorial summary based on the publisher abstract or original publication metadata.

The framework combines logistic-map chaotic key generation, DNA-inspired encoding, and XOR operations to encrypt satellite images. Evaluation considers entropy, adjacent-pixel correlation, and sensitivity measures including NPCR and UACI. The reported statistical indicators are encouraging, but the experiment uses a single grayscale image; larger and more varied evaluations remain necessary.

Authors: Asim Seedahmed Ali Osman, Ibrahim Rizqallah Alzahrani, Aeshah Khalil I Alotaibi, Adil Mahmoud Mohamed Mahmoud, Daifallah Zaid Alotaibe

Editorial summary based on the publisher abstract or original publication metadata.

This review examines 2015–2023 research on deep learning for blood-image analysis and acute lymphoblastic leukemia classification. It compares CNNs, hybrid CNN–LSTM/GRU models, and transfer-learning approaches using precision, recall, F1, and ROC-AUC. The discussion highlights CNN–GRU combinations and challenges involving limited or imbalanced data and resource-constrained computing environments.

Authors: Osama MA Yousif, Eltyeb Elsamani Abdelgabar Elsamani, Ibrahim Rizqallah Alzahrani, Adil Mahmoud Mohamed Mahmoud

Summary prepared from the abstract indexed in Google Scholar.

The paper proposes a BiLSTM-based SM-CJ framework for classifying tweets as hate, offensive, or neutral speech. Text cleaning and TF-IDF representation precede contextual sequence modelling. Forward and backward context and long-range dependencies support detection of nuanced language. Evaluation uses accuracy, precision, recall, and F1, with comparisons against existing approaches.

Authors: IR Alzahrani

Editorial summary based on the publisher abstract or original publication metadata.

The AROTL-DFDBA approach distinguishes genuine and manipulated biometric data using transfer learning and Artificial Rabbits Optimization. A modified DarkNet-53 extracts features, with optimization selecting its hyperparameters, followed by a weighted regularized extreme learning machine classifier. Experiments on the evaluated DeepFake dataset report a maximum accuracy of 96.48%.

Authors: Sana Alazwari, Marwa Obayya Jamal Alsamri, Mohammad Alamgeer, Rana Alabdan, Ibrahim Alzahrani, Mohammed Rizwanullah, Azza Elneil Osman

Editorial summary based on the publisher abstract or original publication metadata.

This work combines kernel random forests and Black Hole Optimization for wearable-sensor data fusion, alongside XGBoost classification of echocardiogram images. It targets more efficient heart-disease prediction from combined medical information. The abstract reports testing accuracies of 95.89% for the sensor-fusion dataset and 96.21% for Cleveland data under the study's experimental settings.

Authors: Ala Saleh Alluhaidan, Mashael Maashi, Noha Negm, Shoayee Dlaim Alotaibi, Ibrahim R Alzahrani, Ahmed S Salama

Summary prepared from the abstract indexed in Google Scholar.

The study addresses semantic web-content classification and inappropriate-content filtering. A dendritic neural network extracts useful keywords, type-2 fuzzy ontology assigns content classes, and pigeon swarm optimization tunes network hyperparameters. The approach aims to retain contextual distinctions that conventional filters can miss, including differences between medical and adult material, and is compared with existing methods.

Authors: Mahmoud Ragab, Fatmah Yousef Assiri, Diaa Hamed, Ibrahim R Alzahrani, Turki Althaqafi, Hadi Oqaibi

Editorial summary based on the publisher abstract or original publication metadata.

The Secure Semantic Communication Framework combines AES encryption with Deep Q-Network-based adversarial training for IoT communications. It seeks to preserve meaning and communication accuracy while limiting disclosure to eavesdroppers. Experimental evaluation reports improved resistance to semantic-data reconstruction from intercepted messages while maintaining communication performance in encrypted and unencrypted modes.

Authors: Ibrahim R Alzahrani

Editorial summary based on the publisher abstract or original publication metadata.

The AOAEL-CDC framework detects and classifies attacks in IoT environments. Artificial Orca Algorithm feature selection is combined with an ensemble of BiLSTM, GRU, and extreme learning machine models. Marine Predators Algorithm tunes model hyperparameters. Benchmark-data simulations evaluate the resulting approach for identifying malicious network activity.

Authors: Randa Allafi, Ibrahim R Alzahrani

Editorial summary based on the publisher abstract or original publication metadata.

The CP3-BSOADL framework predicts consumer purchasing power using behaviour, preference, and contextual information associated with wireless sensor networks. A stacked autoencoder performs prediction, while Barracuda Swarm Optimization tunes its hyperparameters. Evaluation on an e-commerce dataset compares predictive performance with recent methods, with potential uses in personalized recommendations and interest-based commerce.

Authors: Latifah Almuqren, Nuha Alruwais, Asma A Alhashmi, Ibrahim R Alzahrani, Nahla Salih, Mohammed Assiri, K Shankar

Editorial summary based on the publisher abstract or original publication metadata.

This research combines the Analytic Hierarchy Process with support vector machines and decision trees to map coastal flood susceptibility. It examines conditioning factors and RCP 2.6 and RCP 8.5 climate scenarios, including a Bandar Abbas case study. The reported AHP-DT model achieved ROC-AUC of 0.95, informing prioritization of flood-risk reduction measures.

Authors: Mashael M Asiri, Ghadah Aldehim, Nuha Alruwais, Randa Allafi, Ibrahim Alzahrani, Amal M Nouri, Mohammed Assiri, Noura Abdelaziz Ahmed

Editorial summary based on the publisher abstract or original publication metadata.

EBSAEDL-RD targets ransomware recognition and classification in IoT security. The pipeline applies min–max normalization, Ebola Optimization Search for feature selection, a bidirectional GRU classifier, and Sparrow Search for hyperparameter tuning. Benchmark experiments compare the combined method with other approaches, reporting improved performance within the evaluated conditions.

Authors: Ibrahim R Alzahrani, Randa Allafi

Editorial summary based on the publisher abstract or original publication metadata.

GJODL-CADC combines Golden Jackal Optimization feature selection with a hybrid autoencoder–deep belief network for cyberattack detection in industrial IoT. Pelican Optimization adjusts the classifier's parameters. Simulation-based comparisons report promising detection and classification performance against existing approaches within the study's evaluation settings.

Authors: Louai A Maghrabi, Ibrahim R Alzahrani, Dheyaaldin Alsalman, Zenah Mahmoud AlKubaisy, Diaa Hamed, Mahmoud Ragab

Editorial summary based on the publisher abstract or original publication metadata.

The proposed pipeline detects intracranial hemorrhage in CT images using bilateral filtering, Faster SqueezeNet feature extraction, and a denoising autoencoder tuned by Political Optimizer. Benchmark evaluation reports a maximum detection accuracy of 98.43%. The contribution concerns research-stage computer-aided image classification and does not by itself establish prospective clinical effectiveness.

Authors: Mahmoud Ragab, Reda Salama, Fahd S Alotaibi, Hesham A Abdushkour, Ibrahim R Alzahrani

Editorial summary based on the publisher abstract or original publication metadata.

MANFIS-DM combines UAV-network clustering and remote-sensing image classification for autonomous decision support. Quantum differential evolution organizes clusters, while DenseNet feature extraction, Adadelta tuning, and adaptive neuro-fuzzy inference classify images. The work evaluates the combined processes using multiple measures, linking energy-conscious network organization with image-based decision making.

Authors: Mahmoud Ragab, Ehab Bahaudien Ashary, Wajdi H Aljedaibi, Ibrahim R Alzahrani, Anil Kumar, Deepak Gupta, Romany F Mansour

Editorial summary based on the publisher abstract or original publication metadata.

This framework combines cloud IoT and smart data mining for bank-marketing decisions. Six algorithms—Naïve Bayes, logistic regression, random forests, SVM, ID3, and C4.5—are compared on real marketing data using predictive metrics and runtime. C4.5 performed best in the reported comparison, while platform evaluation considered response time, request rate, and data processing.

Authors: Yuan Liu, Ibrahim R Alzahrani, Refed Adnan Jaleel, Saleh Al Sulaie

Editorial summary based on the publisher abstract or original publication metadata.

The MDL-CADDR pipeline grades diabetic retinopathy from fundus images. It combines preprocessing, Archimedes Optimization with Kapur entropy for segmentation, Chimp-optimized DenseNet feature extraction, and spiking neural network classification. Experiments on the MESSIDOR benchmark report a maximum accuracy of 99.73%, describing performance under the paper's specific experimental conditions.

Authors: Mahmoud Ragab, Wajdi H Aljedaibi, Alaa F Nahhas, Ibrahim R Alzahrani

Editorial summary based on the publisher abstract or original publication metadata.

This framework identifies abnormal events in critical industrial infrastructures for digital forensic investigation. Manta Ray Foraging Optimization selects features, an optimized stacked autoencoder classifies events, and Coyote Optimization tunes parameters. The workflow covers data collection, preparation, feature selection, and classification, with comparative validation on benchmark data.

Authors: Abdullah S. AL-Malaise AL-Ghamdi, Mahmoud Ragab, F. J. Alsolami, Hani Choudhry, Ibrahim Rizqallah Alzahrani

Editorial summary based on the publisher abstract or original publication metadata.

The study evaluates a game-based mobile application supporting Arabic reading among primary-school pupils with dyslexia in Saudi Arabia. Forty-five children aged 8–10 participated. The reported analysis found improved reading skills and positive responses to the application, highlighting the value of interaction design aligned with learners' needs in this particular study group.

Authors: Randa Allafi, Ibrahim Alzahrani, Paul Newbury

Editorial summary based on the publisher abstract or original publication metadata.

The paper proposes a layered architecture for healthcare IoMT workflows across mobile, edge, and cloud nodes. Its SEOS framework combines deadline division, task sequencing, homomorphic security, initial scheduling, and variable-neighbourhood search. Evaluation reports lower application makespan than the compared baselines while accommodating workloads with different security and delay requirements.

Authors: Abdullah Lakhan, Qurat-ul-ain Mastoi, Mazhar Ali Dootio, Fehaid Alqahtani, Ibrahim R Alzahrani, Fatmah Baothman, Syed Yaseen Shah, Syed Aziz Shah, Nadeem Anjum, Qammer Hussain Abbasi, Muhammad Saddam Khokhar

Editorial summary based on the publisher abstract or original publication metadata.

This thesis investigates smoother and higher-quality HTTP adaptive video streaming. Its testbed examines segment durations from 1 to 20 seconds, client-buffer resizing, and representation selection under constrained bandwidth. The work studies trade-offs between perceived image quality and playback continuity and explores smart adaptation and machine-learning solutions for improving the viewing experience.

Authors: Ibrahim Rizqallah Alzahrani

Summary prepared from the abstract indexed in Google Scholar.

This paper experimentally investigates DASH segment duration over a local-area network. It adapts client-buffer size to segment length and chooses video representations according to available bandwidth. Tests compare 1, 2, 5, 10, 15, and 20-second segments using objective quality evaluation to examine the relationship between segmentation, buffering, and streaming performance.

Authors: Ibrahim Rizqallah Alzahrani, Naeem Ramzan, Abbes Amira

Summary prepared from the abstract indexed in Google Scholar.

The paper proposes supervised machine learning for selecting video quality and bitrate in MPEG-DASH. Decisions consider bandwidth availability, client-buffer state, and each segment's bitrate. Experimental evaluation reports 79% accuracy in predicting optimal target quality, illustrating the potential of learned adaptation for improving representation requests and quality of experience.

Authors: Ibrahim Rizqallah Alzahrani, Naeem Ramzan, Stamos Katsigiannis, Abbes Amira

Editorial summary based on the publisher abstract or original publication metadata.

This SSRN manuscript examines Twitter attitudes toward the Biden administration's climate policies during its first year. It analyses 5,467 tweets collected with Tweepy, using preprocessing, sentiment analysis, and statistical tests including t-tests. It discusses social-media data as a tool for monitoring public opinion. The record is identified as a manuscript without assuming peer review or an unverified publication date.

Authors: Ibrahim Alzahrani

Summary prepared from the abstract indexed in Google Scholar.