Event Series Anomaly Detection Using Point Process Training
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for supervised abnormality detection in event data are inadequate, as they typically deal with accumulated feature values and cannot accurately detect abnormalities in event data.
Innovation Solution
A model training process that optimizes an objective function representing the relationship between the probability of event occurrence at each time point and the degree of abnormality, using a point process to model past events, allowing for accurate detection of abnormality in event data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If existing supervised abnormality detection methods are used, then the detection process can be performed with available methods, but the detection accuracy is insufficient because these methods deal with accumulated feature values and cannot be applied to event data
Solution Approach 1:
The patent transforms the approach by changing the fundamental parameter representation from accumulated feature values to event series data. It introduces a point process model that directly processes event timestamps and characteristics, thereby adapting the detection system to work with event data while maintaining high detection accuracy through specialized probability density function estimation.
2Measurement precision
If a point process model is used to model past events, then the relationship between event occurrence probability and abnormality degree can be captured, but the model complexity increases
Solution Approach 1:
The patent introduces a probability density function as an intermediary between the raw event series data and the abnormality detection process. This intermediary component (the point process model with PDF estimation) bridges the gap between complex event data and the detection objective, managing model complexity by providing a structured probabilistic framework that systematically processes event sequences.
Data Source
AI summary
An object is to make it possible to accurately detect abnormality of event data.A training unit (105) trains a parameter of a model based on a plurality of event series that are event data in a time series and labels that indicate abnormality or normality with respect to event data of each of the plurality of event series, the model outputting a degree of abnormality of a target event series when the target event series is input, the target event series being an event series of which the degree of abnormality is to be predicted, the parameter being trained to optimize an objective function that represents a relationship between a probability of occurrence of an event at each time point in the time series and a degree of abnormality of each of the plurality of event series.