Loop-Aware Footprint Matrix for Transparent Next-Event Prediction

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing solutions for predicting individual interactions, such as those in employee-focused, user-focused, or customer-focused industries, provide only aggregated analysis and information after the fact, failing to anticipate next events, which is crucial for retention and intervention planning.

Innovation Solution

A computer-implemented method using a loop-aware footprint matrix (LaFM) data structure to store and predict next events by initializing a matrix with data process traces, recording operator functions, and generating path predictions through a process tree data structure, incorporating machine learning and clustering for improved accuracy.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Quantity of substance

If aggregated analysis of many users is provided, then statistical analysis coverage is improved, but individual prediction accuracy deteriorates

Engineering Contradiction:

Improvenumber of users analyzedVSAvoidindividual prediction accuracy

Core Design Contradiction:

Quantity of substanceVSMeasurement precision

Solution Approach 1:

The patent segments the analysis by creating separate footprint matrices for different user clusters identified through machine learning clustering algorithms. Each cluster receives individualized path prediction analysis, transforming the aggregated approach into segmented individualized analysis while maintaining scalability across many users.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The system dynamically adapts the analysis approach by using machine learning to identify user clusters and automatically adjust the footprint matrix construction accordingly. This dynamic segmentation allows the system to provide individualized predictions for each cluster while processing large numbers of users efficiently.

Inventive Principle:

Principle #15Dynamics

2Loss of information

If process mining is used to discover process models, then process understanding is improved, but prediction capability deteriorates

Engineering Contradiction:

Improveprocess understandingVSAvoidprediction timing

Core Design Contradiction:

Loss of informationVSLoss of time

Solution Approach 1:

The patent performs preliminary process discovery to build the footprint matrix data structure in advance, storing execution indices and operator function mappings. This pre-computed structure enables rapid path predictions for individual users without requiring time-consuming process analysis at prediction time, thus maintaining both process understanding and prediction timing.

Inventive Principle:

Principle #10Preliminary action

3Measurement precision

If complex machine learning clustering is applied, then prediction accuracy is improved, but system complexity deteriorates

Engineering Contradiction:

Improveprediction accuracyVSAvoidsystem complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent introduces the footprint matrix as an intermediary data structure that bridges complex machine learning clustering operations and simple path prediction queries. The matrix pre-computes and stores execution indices during clustering, transforming complex analytical work into simple lookup operations during prediction, thus reducing system complexity at query time while maintaining high accuracy.

Inventive Principle:

Principle #24Intermediary (Mediator)

4Reliability

If individual path predictions are generated, then retention planning is improved, but computational resources deteriorate

Engineering Contradiction:

Improveretention planningVSAvoidcomputational resources

Core Design Contradiction:

ReliabilityVSUse of energy by moving object

Solution Approach 1:

The system performs preliminary computation by building the footprint matrix with pre-computed execution indices during the data processing phase. This allows individual path predictions to be generated rapidly using simple matrix lookups rather than complex computations, significantly reducing computational resources required during actual prediction operations while maintaining reliable individualized retention planning.

Inventive Principle:

Principle #10Preliminary action

Data Source

PatentUS12499397B2Accurate and transparent path prediction using process mining

Publication Date: 2025.12.16 ODAIA INTELLIGENCE INC

AI summary

The present disclosure generally relates to the field of data structures and in particular, a loop-aware footprint matrix data structure adapted for data process traversal. The proposed approach is directed to a computer-based analytic system and corresponding method that uses a specific data structure and processing thereof, in some embodiments, adapted to computationally estimate predictions of next events by first generating a data structure based on business process models obtained using process mining techniques, and then using the improved data structure for generating predictions, which can then be encapsulated in the form of computer instructions or machine instruction sets, having a specific sequence for execution.