Inbunden, Engelska, 2027
1 769 kr
Kommande
Integrate anomaly detection algorithms into real-world engineering and data science Detecting anomalies across images, video, and serial data demands fluency in both statistical foundations and modern deep learning architectures. Anomaly Detection unifies classical machine learning, predictive models, representation learning, and ensemble/hybrid designs in a single reference. It spans supervised, semi-supervised, and unsupervised data regimes while addressing evaluation under class skew and deployment considerations. The authors provide case studies drawn from cybersecurity, fraud detection, environmental monitoring, and healthcare, with each study that pairs domain-specific datasets with evaluation metrics and benchmarking protocols. The book also covers: Deep generative models including variational autoencoders and generative adversarial networks applied to image and video anomaly detection tasksEvaluation benchmarking strategies that account for class imbalance, skewed distributions, and domain-specific performance criteria across multiple data modalitiesHybrid and ensemble model designs combining classical statistical methods with deep learning to improve robustness in complex detection scenariosDeployment considerations for transitioning anomaly detection models from research prototypes into scalable, production-ready engineering systemsEngineers, data scientists, and researchers working in cybersecurity, manufacturing, healthcare, and predictive maintenance will find actionable methods for building and evaluating anomaly detection systems. Senior undergraduate and graduate students in data science and analytics gain a structured reference connecting algorithmic foundations to applied workflows.