Distributed Anomaly Detection With Combinable Statistical Summaries
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Maintaining historical measurement information for anomaly detection requires substantial storage space and network bandwidth, especially in devices with limited resources, and obtaining information from other devices is inefficient.
Innovation Solution
Implementing local measurement value summaries that maintain a summarization of measurement values without increasing size, combined with a central computing device to create a global summary for anomaly detection, allowing devices to identify outliers efficiently.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If historical measurement information is maintained for anomaly detection, then anomaly detection capability is improved, but storage space requirement increases
Solution Approach 1:
The patent extracts only the essential statistical characteristics (mean, variance, skewness, kurtosis) from the full historical measurement data, discarding the raw data itself. This allows anomaly detection capability to be maintained while significantly reducing storage requirements, as only these four parameters need to be stored rather than complete measurement histories.
Solution Approach 2:
The patent creates a simplified copy of the measurement distribution using statistical moments instead of storing the actual measurement data. This copy contains sufficient information for anomaly detection while occupying minimal storage space, effectively replacing the need to store complete historical records.
2Reliability
If measurement histories from multiple devices are obtained, then anomaly detection accuracy is improved, but network bandwidth consumption increases
Solution Approach 1:
The patent extracts only the four statistical moments from each device's measurement history and transmits only these compressed values across the network. This dramatically reduces network bandwidth consumption compared to transmitting complete measurement histories, while still enabling accurate cross-device anomaly detection through the combined statistical profile.
3Measurement precision
If complete measurement histories are stored, then anomaly detection precision is improved, but device complexity increases
Solution Approach 1:
The patent extracts only the necessary statistical parameters (mean, variance, skewness, kurtosis) from the measurement data, eliminating the need to manage and store complete measurement histories. This simplifies data management complexity significantly while maintaining anomaly detection precision, as the four statistical moments provide sufficient information for accurate anomaly identification.
Data Source
AI summary
Distributed anomaly detection using combinable measurement value summaries is disclosed. A computing device obtains a global combined measurement value summary. The global combined measurement value summary includes a combined measurement value summary of a plurality of measurement values taken by a plurality of other computing devices measuring a first type of item. The computing device obtains a measurement value of an item of the first type. It is determined whether the new measurement value is an anomalous measurement value based at least partially on the global combined measurement value summary.