New method uses difference graphs for root cause analysis in time-series data

Researchers have developed a new method for root cause analysis in linear time-series data, focusing on identifying variables with changing causal coefficients during anomalies. This approach, termed difference graph discovery, adapts techniques for comparing two populations to a time-series context, distinguishing between normal and anomalous operational regimes. The method was evaluated on simulated data and demonstrated effectiveness on real-world datasets from IT monitoring and intensive care, aiding in the localization of causal mechanisms behind anomalous system behavior. AI

IMPACT This research could improve anomaly detection and diagnosis in complex systems by providing a more precise way to identify the underlying causes.

RANK_REASON The item is an academic paper published on arXiv detailing a new method for root cause analysis. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv cs.AI →

• alphaXiv
• arXiv
• arXivLabs
• CatalyzeX Code Finder for Papers
• Connected Papers
• CORE Recommender
• DagsHub
• difference graph discovery
• Dynamic Structural Causal Models
• Gotit.pub
• Hugging Face
• Influence Flower
• Intensive care monitoring
• linear time-series data
• Litmaps
• root cause analysis
• ScienceCast
• scite Smart Citations

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →