Data-Agnostic Anomaly Detection via Normalcy Bounds
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current anomaly detection techniques in complex systems are limited by their reliance on parametric methods, which assume known data distributions, and non-parametric methods are resource-intensive and complex, making them unsuitable for diverse data types and leading to false positives.
Innovation Solution
A data-agnostic computational system that aggregates data, performs quality assurance, and categorizes it using routines like parametric category detection, data density detection, and variability detection to establish normalcy bounds, enabling accurate anomaly detection in various complex systems without assuming data distributions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If parametric anomaly detection techniques are used, then detection accuracy is improved for specific data types, but applicability to diverse data types deteriorates
Solution Approach 1:
The patent implements a universal anomaly detection system that automatically adapts to different data types through data-agnostic procedures. The system categorizes data into four types (periodic, non-periodic, high-variability, low-variability) and applies appropriate detection methods for each, enabling a single system to handle diverse data types effectively without requiring manual configuration or assumption of data distributions.
2Adaptability or versatility
If non-parametric anomaly detection techniques are used, then applicability to diverse data types is improved, but computational complexity and resource consumption increase
Solution Approach 1:
The patent segments the anomaly detection process into distinct procedural branches based on data categorization. By dividing data into four categories and applying specialized detection procedures to each segment, the system achieves comprehensive coverage of diverse data types while maintaining computational efficiency through targeted rather than universal complex algorithms.
3Reliability
If traditional anomaly detection methods are used, then detection capability is improved, but false positive rate increases
Solution Approach 1:
The patent dynamically adjusts detection parameters and procedures based on the categorized characteristics of the input data. By changing the detection approach according to data type (periodic, non-periodic, high-variability, low-variability), the system optimizes detection sensitivity for each data type, thereby reducing false positives while maintaining high anomaly detection capability.
Data Source
AI summary
This disclosure presents computational systems and methods for detecting anomalies in data output from any type of monitoring tool. The data is aggregated and sent to an alerting system for abnormality detection via comparison with normalcy bounds. The anomaly detection methods are performed by construction of normalcy bounds of the data based on the past behavior of the data output from the monitoring tool. The methods use data quality assurance and data categorization processes that allow choosing a correct procedure for determination of the normalcy bounds. The methods are completely data agnostic, and as a result, can also be used to detect abnormalities in time series data associated with any complex system.