Cross-Temporal Anomaly Detection Using Event Record Profiles
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing anomaly detection systems fail to effectively process temporally dynamic input data, lacking in both temporally-aware and non-temporally-aware machine learning models, which leads to inefficiencies in predictive anomaly detection across time.
Innovation Solution
The proposed solution preprocesses input data using cross-temporal predictive inferences that leverage intra-period and inter-period relationships of event records, then processes these preprocessed data using an anomaly detection machine learning model to generate temporally-aware anomaly detections, employing computationally resource-efficient models that provide probabilistic and interpretable outputs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing anomaly detection systems process temporally dynamic input data without temporally-aware machine learning models, then the processing speed is maintained, but the detection accuracy and reliability across time periods deteriorates
Solution Approach 1:
The patent segments the anomaly detection process into distinct temporal components: intra-period processing (within each time period) and inter-period processing (across time periods). This segmentation allows the system to handle temporal dynamics systematically by creating separate processing pathways for different temporal scales, improving reliability without overwhelming system complexity
Solution Approach 2:
The patent introduces a temporal dimension to the anomaly detection system by processing data across multiple time periods. It adds intra-period and inter-period processing layers that operate in the time dimension, transforming the detection approach from static to temporal-aware, thereby improving reliability across time while managing complexity through structured dimensional expansion
2Productivity
If computationally resource-efficient models are used for anomaly detection, then processing efficiency is improved, but the ability to capture complex temporal relationships deteriorates
Solution Approach 1:
The patent segments temporal relationship capture into two efficient components: intra-period relationships (within each time period) and inter-period relationships (across time periods). This segmentation allows computationally efficient processing by handling temporal dependencies in modular stages rather than requiring complex global temporal models, maintaining productivity while improving temporally-aware detection capability
Solution Approach 2:
The patent performs preliminary processing of temporal relationships by pre-computing intra-period and inter-period patterns before final anomaly detection. This preliminary action extracts temporal features in advance, enabling efficient detection models to work with pre-processed temporal information, thereby maintaining productivity while capturing complex temporal relationships
3Measurement precision
If cross-temporal predictive inferences are performed using intra-period and inter-period relationships, then anomaly detection accuracy is improved, but computational complexity increases
Solution Approach 1:
The patent segments cross-temporal predictive inference into intra-period inference (within each time period) and inter-period inference (across time periods). This segmentation breaks down complex cross-temporal analysis into manageable components, improving measurement precision by systematically capturing temporal patterns while controlling processing complexity through structured decomposition
Solution Approach 2:
The patent introduces event record profiles as intermediary structures that mediate between raw event data and anomaly detection. These profiles aggregate and structure temporal relationships, serving as intermediaries that simplify complex cross-temporal computations while preserving detection precision, thereby improving measurement precision without proportionally increasing processing complexity
4Loss of information
If event record profiles are generated to describe temporally-related event code data objects, then interpretability is improved, but data processing time increases
Solution Approach 1:
The patent extracts essential temporal information from event records to create compact event record profiles. By taking out only the most relevant temporal features and relationships needed for anomaly detection, the system improves information interpretability while minimizing the time cost of profile generation, avoiding processing redundancy
Data Source
AI summary
There is a need for more effective and efficient predictive anomaly detection. This need can be addressed by, for example, solutions for performing anomaly detection using an anomaly detection machine learning model. In one example, a method includes: identifying a plurality of event records; for each event record of the plurality of event records, determining a temporally-related event code data object based at least in part on a temporally-related subset of the one or more event codes that is associated with the event record; generating one or more event record profiles based on each temporally-related event code data object; processing the one or more event record profiles using an anomaly detection machine learning model to generate one or more anomaly detection predictions; and performing one or more prediction-based actions based at least in part on the one or more anomaly detections.