Exploring the impact of Optimised Hyperparameters on Bi-LSTM-based Contextual Anomaly Detector

2025/01/25 by Aafan Ahmad Toor, Toor, Aafan Ahmad, Jia‐Chun Lin +3
Computer Science · #Anomaly Detection Techniques and Applications #Artificial Intelligence (cs.AI) #FOS: Computer and information sciences #Machine Learning (cs.LG) #Network Security and Intrusion Detection

paper · pdf · doi:10.48550/arxiv.2501.15053

openalex publication_date 2025/01/25 · openalex created_date 2025/01/29 · openalex updated_date 2026/07/28

Abstract

Smart Home Repository of Time series with Contextual Anomalies (SHoRT-CA) comprises multivariate time series data measuring indoor air quality in a smart home. The repository comprises multiple datasets, including data values of air quality, human activity, and environmental interaction, which contain data recorded at different time intervals and periods. The real-world contextual anomalies are induced into the datasets by performing various experiments to affect the sensors. The repository consists of data files, anomaly descriptions, and figures of sensor values and highlighted anomalies.

Related

• SALAD: Self-Adaptive Lightweight Anomaly Detection for Real-time Recurrent Time Series
• Are we certain it's anomalous?
• RePAD: Real-time Proactive Anomaly Detection for Time Series
• Artificial Intelligence-Based Multiscale Temporal Modeling for Anomaly Detection in Cloud Services
• LSTM-Autoencoder based Anomaly Detection for Indoor Air Quality Time Series Data