ML Aggregation Code Generation for Precomputed Transaction Data

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing machine learning models require resource-intensive real-time aggregation data calculation and are not easily adaptable across different geographic and demographic data sets, leading to the creation of numerous models and extensive data requirements.

Innovation Solution

A system and method for generating code to retrieve aggregation data for machine learning models, using an XML data file to create a code generation template with transaction aggregate parameters, enabling efficient retrieval of data from databases.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If aggregation data is calculated in real-time for machine learning models, then the model can process current data accurately, but the resource intensity and time required increase significantly

Engineering Contradiction:

Improvedata aggregation accuracyVSAvoidresource intensity

Core Design Contradiction:

Measurement precisionVSUse of energy by moving object

Solution Approach 1:

The system pre-calculates and stores aggregation data in a database before it is needed for machine learning model processing. This preliminary action eliminates the need for real-time calculation, significantly reducing resource intensity while maintaining data accuracy. The aggregation data is computed in advance and stored for later retrieval by the machine learning model.

Inventive Principle:

Principle #10Preliminary action

2Measurement precision

If aggregation data is calculated in real-time, then the model reflects current data state, but the time required for calculation increases

Engineering Contradiction:

Improvedata aggregation accuracyVSAvoidcalculation time

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system performs aggregation data calculation in advance and stores the results in a database. This preliminary computation eliminates real-time calculation delays, allowing the machine learning model to quickly retrieve pre-computed data without time-consuming calculations during processing.

Inventive Principle:

Principle #10Preliminary action

3Measurement precision

If machine learning models are developed for specific geographic areas and demographics, then the model accuracy for that specific data is improved, but the number of models and data requirements increase

Engineering Contradiction:

Improvemodel accuracyVSAvoidnumber of models

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The system creates a single machine learning model that can process multiple types of aggregation data from different geographic areas and demographics. The model is designed to be universal and can handle various data types by retrieving appropriate aggregation data from the database, eliminating the need to create separate models for each geographic area or demographic group while maintaining high accuracy.

Inventive Principle:

Principle #6Universality (Multi-functionality)

4Measurement precision

If multiple machine learning models are created for different data sets, then each model can be optimized for its specific data, but the overall system complexity and data storage requirements increase

Engineering Contradiction:

Improvemodel optimizationVSAvoiddata volume

Core Design Contradiction:

Measurement precisionVSQuantity of substance

Solution Approach 1:

The system implements a single machine learning model that can process multiple types of aggregation data from different sources. Instead of creating multiple specialized models, the universal model retrieves and processes various aggregation data types from a centralized database, reducing overall data storage requirements while maintaining optimization for different data sets.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Solution Approach 2:

The system merges multiple data sets and aggregation data types into a single database structure that can be queried by one machine learning model. This consolidation combines what would otherwise require multiple separate models and data storage systems into a unified approach, reducing overall system complexity and data volume requirements.

Inventive Principle:

Principle #5Merging (Combining)

Data Source

PatentUS12554708B2System, method, and computer program product for generating code to retrieve aggregation data for machine learning models

Publication Date: 2026.02.17 VISA INTERNATIONAL SERVICE ASSOCIATION

AI summary

Provided is a system that includes at least one processor programmed or configured to receive an XML data file, wherein the XML data file includes data associated with one or more input parameters of a machine learning model, generate a code generation template based on the data associated with one or more input parameters of the machine learning model included in the XML file, where the code generation template includes one or more keys associated with one or more parameters of a transaction aggregate for an account of a user, and generate a file of executable code based on the code generation template, wherein the file of executable code includes instructions that, when executed by at least one processor, causes at least one processor to retrieve transaction aggregate data associated with the transaction aggregate for the account of the user. A method and computer program product are also provided.