PeerJ

Deep learning has become a central paradigm in modern computer science, with convolutional neural networks (CNNs) and Transformer-based architectures representing two of its most influential model families. CNNs have historically dominated visual recognition tasks since the ImageNet breakthrough, leveraging local connectivity and weight sharing to achieve parameter-efficient representations and favorable computational scaling (Krizhevsky, Sutskever & Hinton, 2012).

Explainable fraud detection under fixed false positive rate Duluth, Georgia, United States DOI 10.7717/peerj-cs.4007 Academic Editor Ankit Vishnoi Subject Areas Algorithms and Analysis of Algorithms, Artificial Intelligence, Data Mining and Machine Learning, Data Science, Neural Networks Keywords Fraud detection, Explainable artificial intelligence (XAI), Machine learning, Imbalanced classification, False positive rate, Operational machine learning, SHAP, Model interpretability, Risk scoring,...

Escape response and perch-site choice in shrikes: effects of human disturbance in Europe and Southeast Asia 1Faculty of Sciences, University of Siedlce, Siedlce, Poland 2Department of Vertebrate Ecology and Paleontology, Institute of Environmental Biology, University of Environmental and Life Sciences, Wrocław, Poland 3Faculty of Environmental Sciences, National University of Laos, Vientiane, Laos DOI 10.7717/peerj.21506 Academic Editor Javier Manjarrez Subject Areas Animal Behavior, Ecology,...

This study systematically evaluated the effects of brisk walking on blood pressure using a Bayesian dose–response meta-analytic framework. The findings indicate that brisk walking significantly reduced both systolic and diastolic blood pressure, with effect sizes of moderate magnitude. A “U” shaped dose–response relationship was observed between brisk walking and systolic blood pressure, whereas a monotonic negative dose–response relationship was identified for diastolic blood pressure.

The results of this study suggest that light intensity and plant density interact under the specific conditions of the experiment to influence sweet pepper yield, particularly under low natural light conditions during winter months.

The current study aimed to evaluate near vision-related symptoms and their associated visual functions among medical sciences students in Qassim University, Saudi Arabia. The results show a high occurrence of near-vision symptoms in this population, where 29.3% of participants were classified as symptomatic according to the CISS.

1Laboratoire de Neurosciences Cognitives et Adaptatives, Université de Strasbourg, Strasbourg, France 2Centre de Primatologie, Université de Strasbourg, Niederhausbergen, France DOI 10.7717/peerj.21377 Academic Editor Jennifer Vonk Subject Areas Animal Behavior, Zoology Keywords Intuitive statistics, Proportional reasoning, Probabilistic inference, Decision-making, Numerical cognition, Non-human primates Copyright © Licence This is an open access article distributed under the terms of the...

Xinjiang Medical University, Urumqi, Xinjiang, China DOI 10.7717/peerj.21372 Academic Editor Lesley Anson Subject Areas Gastroenterology and Hepatology, Immunology, Respiratory Medicine Keywords Bronchial asthma, Mouse model, Inflammation, Ulcerative Colitis Copyright © Licence This is an open access article distributed under the terms of the Creative Commons Attribution License, which permits using, remixing, and building upon the work non-commercially, as long as it is properly attributed.

The article studied short-term traffic flow prediction on large-scale road networks, a fundamental task in intelligent transportation systems and urban traffic management. As traffic environments become increasingly complex and heterogeneous, accurate and robust forecasting is crucial for congestion mitigation, route planning, and real-time control.

A blockchain-enabled hybrid intrusion detection framework for resource-constrained devices School of Electrical Engineering and Computer Science, University of North Dakota, Grand Forks, North Dakota, United States DOI 10.7717/peerj-cs.3899 Academic Editor Rowayda Sadek Subject Areas Computer Networks and Communications, Cryptography, Security and Privacy, Internet of Things, Blockchain Keywords Intrusion detection system, Internet of Things, Federated learning, Blockchain, Edge computing...