Model-Based Query Assignment for Mortgage Loan Evaluation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing systems in the secondary mortgage market are inefficient in processing large numbers of mortgage loans quickly due to their reliance on limited, static data from professional appraisers, leading to a labor-intensive and slow evaluation process.
Innovation Solution
The implementation of systems, methods, and computer products that aggregate and process dynamic data from databases to instantiate models for value, condition, and fraud data, allowing for the efficient assignment of queries to streamlined processes, thereby reducing the need for human appraisal and speeding up the evaluation of mortgage loans.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If systems rely on limited static data from professional appraisers to generate appraisals and evaluate risks, then measurement precision may be maintained, but productivity deteriorates due to labor-intensive processing of large numbers of loans
Solution Approach 1:
The patent segments the appraisal and risk evaluation process into multiple independent components: automated valuation models (AVMs) that process property data, fraud detection models that analyze transaction patterns, and condition assessment models that evaluate property status. Each model operates independently on specific data types, allowing parallel processing of mortgage loans while maintaining comprehensive analysis through the aggregation of multiple specialized assessments.
2Productivity
If systems process large numbers of mortgage loans quickly, then productivity improves, but measurement precision deteriorates due to reliance on automated rather than professional appraisal systems
Solution Approach 1:
The patent merges multiple automated assessment models (valuation, fraud detection, and condition evaluation) into a comprehensive risk evaluation system. By combining the outputs of these specialized models, the system achieves both high processing volume through automation and maintained precision through multi-faceted analysis that mirrors the thoroughness of professional appraisals.
Solution Approach 2:
The system incorporates feedback mechanisms where the results from automated models are continuously refined based on actual loan performance data and professional appraisal comparisons. This feedback loop allows the automated system to improve its measurement precision over time while maintaining high processing speeds, as the models learn from and adjust to real-world outcomes.
3Measurement precision
If systems use comprehensive dynamic data from databases, then measurement precision improves through better risk assessment, but device complexity increases due to multiple data structures and models
Solution Approach 1:
The patent segments the complex data processing system into distinct, modular components: separate data structures for property information, transaction history, and market data; independent valuation models, fraud detection models, and condition assessment models. This segmentation allows each component to be developed, maintained, and optimized independently, reducing overall system complexity while enabling comprehensive risk evaluation through their coordinated operation.
Data Source
AI summary
A system is disclosed that includes a database, a processor in communication with the database, and a memory device in communication with the processor. The memory device stores instructions that, when executed by the processor, perform operations including maintaining at the database a data structure comprising historical value data and historical condition data and receiving a query requesting a streamlined process, where the query comprises submitted value data. The operations further include, based on the query, instantiating a value model by obtaining the historical value data from the database and determining, based on the value data, a modeled value ratio; determining, based on the submitted value data, a submitted value ratio; making a first assessment whether the submitted value ratio is within a predetermined range of the modeled value ratio; making a second assessment whether the modeled value ratio exceeds a predetermined maximum; and, based on the first assessment and the second assessment, assigning a first flag to the query. The operations further include, further based on the query, instantiating a condition model by obtaining the historical condition data from the database and, based on the historical condition data, assigning a second flag to the query. The operations still further include, based on the first flag and the second flag, determining whether to assign the query to the streamlined process.