Anomalous Value Detection in Communications Systems
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for anomaly detection in communications networks fail to accurately handle time-varying range estimation and require non-trivial hyperparameter tuning, leading to inefficient anomaly detection and potential unnecessary interventions.
Innovation Solution
A computer-implemented method by a first node that determines anomalous values by defining subsets of equally spaced time periods, calculating second variability, and detecting anomalies based on time-varying variability, allowing for more accurate detection and reduced unnecessary interventions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If standard decomposition algorithms (ARIMA, Holt-Winters, STL) are used to detect anomalies by breaking down time series components, then anomaly detection capability is improved, but hyperparameter tuning complexity increases significantly
Solution Approach 1:
The patent extracts and removes seasonal components from the time series data through seasonal decomposition before applying anomaly detection. This separation allows the anomaly detector to focus only on irregular variations without being confounded by seasonal patterns, thereby improving detection accuracy while reducing the need for complex hyperparameter tuning to account for seasonal effects
Solution Approach 2:
The patent segments the time series data into distinct components (seasonal, trend, and irregular) and processes each separately. By dividing the complex time series analysis into manageable segments, the system achieves high anomaly detection accuracy without requiring the entire system to be optimized through complex hyperparameter tuning
2Productivity
If constant deviation assumption is used in forecasting systems, then computational simplicity is maintained, but time-varying range estimation accuracy deteriorates
Solution Approach 1:
The patent implements dynamic, time-varying range estimation by computing separate parameters for each decomposed time series component (seasonal, trend, irregular) rather than assuming constant deviation. This dynamic approach adapts to changing conditions in the data while maintaining computational efficiency through the use of standardized statistical methods applied to each component
3Measurement precision
If brute-force grid search is used for hyperparameter tuning, then optimal parameters may be found, but computational cost increases significantly
Solution Approach 1:
The patent performs preliminary seasonal decomposition to remove seasonal components before anomaly detection. This preliminary action simplifies the subsequent anomaly detection task, reducing the dimensionality of the parameter space that would otherwise require exhaustive grid search, thereby achieving good parameter optimization without prohibitive computational cost
Data Source
AI summary
A method, performed by a first node. The method is for handling anomalous values. The first node determines whether an anomalous value is present in a first distribution of values (M) over a first time period. The values are indicative of a performance of a communications system. The first distribution has a first variability per time point. The determining including defining a subset of second time periods within the first time period. The second time periods are equally spaced in time. The determining including determining a second variability of a second distribution (S) of a subset of the values corresponding to the subset of second periods. The determining further including detecting the presence of the anomalous value according to a threshold, based on a variation along time of the second variability. The first node also provides a result of the determination.