Ordinal Graphical Event Model for Order-Sensitive Prediction
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods fail to effectively capture the order-sensitive historical dependence of events in multivariate temporal data, which is crucial for predicting agent interactions and understanding causal relationships in dynamic systems.
Innovation Solution
The development of an ordinal graphical event model (OGEM) that learns conditional intensity rates based on the order of preceding events, using a masking function to disregard repeated event labels and applying piece-wise constant rates over time, allowing for the prediction of event occurrences in asynchronous and irregularly spaced event streams.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If traditional event models are used that ignore event order, then the model complexity is reduced, but the prediction accuracy deteriorates because order-sensitive historical dependence is not captured
Solution Approach 1:
The patent segments the historical event sequence into ordered parent-event types and child-event types, creating distinct segments that preserve temporal order. The masking function segments repeated event labels to maintain only the first occurrence, thereby preserving order sensitivity while managing model complexity through structured segmentation of event sequences.
2Loss of information
If the model captures detailed event order information, then the understanding of causal relationships is improved, but the computational resources and data processing requirements increase
Solution Approach 1:
The patent extracts only the essential order-sensitive information from event sequences by identifying parent-event types and their temporal relationships with child events. The masking function extracts and retains only the first occurrence of each event label, removing redundant repeated information while preserving the critical order-dependent causal relationships.
3Adaptability or versatility
If piece-wise constant conditional intensity rates are used with rate changes at order instantiation changes, then the model adapts to dynamic event patterns, but the number of parameters and model complexity increase
Solution Approach 1:
The patent implements dynamic adaptability by allowing conditional intensity rates to change piece-wise constantly whenever the order instantiation of parent events changes. This dynamic structure enables the model to adapt to varying event patterns over time while maintaining a manageable parameter structure through the piece-wise constant approach rather than continuous variation.
Data Source
AI summary
A computer implemented method of modeling agent interactions, includes receiving event occurrence data. One or more parent-event types and one or more corresponding child-event types are learned from the event occurrence data. A timeline of the one or more parent-event types and one or more corresponding child-event types is modeled from the event occurrence data. Agent interactions are predicted based on an order of the parent-event types in a predetermined history window.