Predictive Model Infrastructure for Dynamic Scoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems face inefficiencies in model development, deployment, and execution for direct marketing campaigns due to disparate data sources, language translation requirements, and infrequent scoring, leading to increased time, cost, and obsolete customer scores.
Innovation Solution
An integrated infrastructure and architecture for rapid development and execution of predictive models, including a centralized data management system, data extraction tool, model validation tool, and model execution tool, which enables seamless data standardization, validation, and dynamic scoring using current customer data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If data is stored in multiple disparate sources with unique definitions and access requirements, then data can be collected from various sources, but modelers need to standardize data and create data tables which results in inconsistent results, compliance risks, and low confidence in outcomes
Solution Approach 1:
The system segments data management into distinct components: a centralized data repository that stores standardized data, extraction tools that retrieve data, and validation tools that ensure quality. This segmentation allows versatile data collection while maintaining reliability through standardized processing pipelines.
Solution Approach 2:
The patent introduces intermediary components including a centralized data repository that acts as a mediator between disparate data sources and modeling tools, and validation tools that mediate between raw data and model execution. These intermediaries standardize data formats and ensure compliance, resolving the contradiction between data versatility and result reliability.
2Ease of manufacture
If model logic is implemented in the development phase using one programming language and deployed in another programming language, then models can be developed with specific language features, but converting logic code into system compatible code requires resources with high technical skills which results in numerous errors and prolonged implementations
Solution Approach 1:
The system uses copying by maintaining model logic in its original programming language format and using extraction tools to copy and translate the logic into system-compatible code. This approach preserves the developmental flexibility of the original language while enabling deployment in different environments, reducing both errors and deployment time.
Solution Approach 2:
The extraction tool acts as an intermediary that automatically translates model logic from development programming languages to system-compatible programming languages. This intermediary handling of code conversion eliminates the need for manual translation by skilled resources, reducing errors and deployment time while maintaining development flexibility.
3Stability of the object's composition
If model execution occurs at a mainframe location and is based on billing cycles, then execution follows a structured schedule, but customer billing cycles vary across a month which takes a full month to score a customer base and customer scores may become obsolete
Solution Approach 1:
The system transitions from static billing-cycle-based execution to dynamic on-demand execution. The model execution tool allows models to be scored immediately when needed, regardless of billing cycle timing. This dynamic approach maintains the structured execution framework while enabling real-time scoring, preventing score obsolescence and reducing the one-month delay inherent in billing-cycle-based execution.
4Productivity
If the infrastructure enables rapid development and execution of predictive models with integrated data management, validation, and execution tools, then model deployment time is reduced and customer scores are ever-ready, but the system requires complex integration of multiple tools and processes
Solution Approach 1:
The patent merges multiple previously separate functions into an integrated system: data extraction, validation, and model execution tools are combined into a unified infrastructure. This integration enables rapid model deployment and ever-ready customer scores by eliminating manual handoffs between separate systems, while the modular architecture manages complexity through standardized interfaces between components.
Data Source
AI summary
A system that enables development and execution of predictive models comprises a centralized data management system, a data extraction tool a model validation tool and a model execution tool. In embodiments, a data management system includes a data management server that can be accessed via a web browser that stores data. An extraction tool includes a data filter adapted to filter data based on, for example, a population criteria, a sample size, and a date range criteria. A model validation tool validates the model. A model execution tool allows a user to score the model.