Dynamic Asset Pairing for Real-Time Time Series Anomaly Detection
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Manual anomaly detection processes in trading systems are time-consuming and prone to inaccuracies, failing to detect emerging behaviors and changes in fund compositions that affect NAV calculations.
Innovation Solution
A system utilizing an anomaly detection engine that identifies asset pairs based on time series data correlations, detects anomalies in real-time data streams, and generates control signals to address detected anomalies.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual anomaly detection processes are used, then analysts can review trading activity, but the process is time-consuming and fails to catch all anomalies
Solution Approach 1:
The patent replaces the manual mechanical review process with an automated computer-based system that uses machine learning models and algorithms to detect anomalies in trading activity, thereby eliminating the time-consuming nature of manual review while improving detection accuracy
Solution Approach 2:
The system enables self-service anomaly detection by automatically monitoring trading activity, identifying anomalies, and generating alerts without requiring continuous manual analyst intervention, allowing the system to serve itself in detecting and reporting anomalies
2Adaptability or versatility
If manual review processes are used, then analysts can identify anomalies, but changes in fund behavior over time make identification difficult
Solution Approach 1:
The patent implements dynamic anomaly detection by continuously training and updating machine learning models with recent data, allowing the system to adapt to changing fund behaviors and market conditions over time, thereby maintaining high detection accuracy despite behavioral changes
Solution Approach 2:
The system incorporates feedback mechanisms where detected anomalies and model performance are used to continuously refine and retrain the detection algorithms, enabling the system to learn from past performance and improve its adaptability to new behaviors while maintaining precision
3Productivity
If automated anomaly detection is implemented, then detection speed improves, but system complexity increases
Solution Approach 1:
The patent divides the anomaly detection system into modular components including data collection modules, preprocessing modules, multiple specialized detection algorithms, and output modules, allowing each component to be independently optimized and managed, thereby achieving high speed without overwhelming complexity
Solution Approach 2:
The system employs universal machine learning frameworks and algorithms that can detect multiple types of anomalies across different asset classes and timeframes using the same core technology platform, reducing overall system complexity while maintaining high productivity through multi-functional capabilities
Data Source
AI summary
Systems, methods, and computer-readable storage media facilitating anomaly data detection are disclosed. In the disclosed embodiments, asset pairs may be identified based on relationship information extracted from time series data. A real-time data stream may be monitored in view of the relationship information, and an anomaly in the asset pairs is detected in real time based on behaviors of assets relative to assets in identified asset pairs. Detection of an anomaly may trigger generation of a control signal that may initiation investigation of the anomaly and other actions to mitigate the impact of the anomaly.