Anomaly Detection via Dimensionality Reduction and Heuristic Mapping
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing anomaly detection methods in industrial systems rely solely on statistical methods, leading to a high number of false positives and are not effective for non-real value sensor data types, such as categorical data.
Innovation Solution
The proposed solution employs a data-driven approach combining dimensionality reduction techniques, change point detection algorithms, and heuristic search methods to identify regions of interest and detect anomalies by mapping data to predefined groups representing normal operating conditions, using control limits and heuristic algorithms to reduce false positives.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Device complexity
If statistical methods are used for anomaly detection, then the detection process is simple, but the number of false positives increases
Solution Approach 1:
The patent segments the anomaly detection process into multiple stages: data preprocessing, feature extraction, pattern recognition, and anomaly scoring. By dividing the detection process into distinct modules, the system reduces false positives while maintaining manageable complexity through structured organization of detection functions.
Solution Approach 2:
The patent transforms raw sensor data into multiple derived parameters and features, changing the parameter space from simple statistical values to complex multi-dimensional characteristics. This parameter transformation enables more accurate anomaly detection by capturing subtle patterns that simple statistical methods miss, thereby reducing false positives.
2Measurement precision
If multiple sensors are employed to monitor system state, then real-time monitoring capability is improved, but the amount of data generated increases
Solution Approach 1:
The patent extracts only the most relevant features and parameters from the large volume of sensor data, separating essential information from redundant data. By taking out only the critical features needed for anomaly detection, the system maintains high monitoring precision while significantly reducing the amount of data that needs to be processed and stored.
Solution Approach 2:
The patent transforms high-dimensional raw sensor data into a lower-dimensional feature space through dimensionality reduction techniques. This dimensional transformation preserves the essential monitoring capabilities while reducing data volume, allowing the system to maintain measurement precision with less data to process.
3Ease of manufacture
If existing statistical models are used for anomaly detection, then the implementation is straightforward, but the methods are not effective for categorical data types
Solution Approach 1:
The patent develops a universal anomaly detection framework that can handle multiple data types including continuous, discrete, and categorical data. By creating a multi-functional detection system that adapts to different data types through appropriate transformation and encoding methods, the patent maintains ease of implementation while significantly improving adaptability across various sensor data types.
Data Source
AI summary
The technique relates to a system and method for data-driven anomaly detection. This technique involves identifying region of interest from the data based on dimensionality reduction technique and change point detection algorithm. A reference data can be obtained separately or can be obtained from the test data also, wherein the reference data represent the normal operating condition of a system. The reference data are classified into different groups representing different modes of operation of the system. A control limit is determined for the different groups. The data within the region of interest are mapped with the different groups of the reference data and it is determined if the mapped data fall outside of the control limit of the mapped group. Finally, at least one abnormal event is detected by applying a heuristic algorithm on the data within the region of interest which are outside the control limit.