Revolving Credit Pattern Detection via Trigger Spend Segmentation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing pattern-detection techniques are ineffective in identifying customer-behavior patterns at an account level in aggregated transaction data from revolving-credit accounts, such as credit card accounts, due to offsetting patterns that mask individual spending and balance fluctuations.
Innovation Solution
The system reviews aggregated transaction data from multiple revolving-credit accounts to detect 'trigger spends' and aligns transaction data relative to individual customers' trigger months for time-series analysis, identifying credit-usage patterns like high spend followed by increasing balance and decreasing spend, which indicates poor credit management.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If aggregated transaction data from multiple customers is analyzed using traditional pattern-detection techniques, then large-scale customer-behavior patterns can be identified, but individual account-level patterns are masked due to offsetting effects
Solution Approach 1:
The patent segments the aggregated transaction data by identifying individual accounts that experienced trigger spends, then analyzes each segment separately. This segmentation allows pattern detection at the account level while still utilizing the aggregated data set, resolving the contradiction between analyzing large volumes of data and detecting precise individual patterns.
Solution Approach 2:
The patent extracts specific accounts from the aggregated data that meet the trigger spend criterion (spend >= 20% of credit line). By taking out these specific accounts for separate analysis, the method preserves individual account-level patterns that would otherwise be masked in the aggregated data, while still benefiting from the large data volume.
2Productivity
If traditional pattern-detection is applied to aggregated revolving-credit data, then computational efficiency is maintained, but individual customer spending and balance patterns cannot be detected
Solution Approach 1:
The patent performs preliminary filtering to identify accounts with trigger spends before conducting detailed pattern analysis. This preliminary action reduces the data set to only those accounts likely to show relevant patterns, maintaining computational efficiency while preserving individual account information that would be lost in full aggregated analysis.
Solution Approach 2:
The patent applies different analysis approaches to different segments of the data: aggregated analysis for overall trends and individual account analysis for trigger spend cases. This local quality approach allows computational efficiency for the majority of data while capturing individual patterns where they matter most.
3Measurement precision
If account-level pattern detection is performed on all aggregated data, then individual customer behaviors can be identified, but computational resources and time are excessively consumed
Solution Approach 1:
The patent applies preliminary filtering to identify trigger spend accounts before detailed pattern analysis. This preliminary action dramatically reduces the number of accounts requiring computationally intensive individual analysis, thus maintaining high measurement precision for relevant cases while minimizing time loss.
Solution Approach 2:
The patent applies full account-level analysis only to the subset of accounts that experienced trigger spends, rather than all accounts. This partial action approach concentrates computational resources on cases where individual pattern detection is most valuable, reducing overall processing time while maintaining precision where needed.
Data Source
AI summary
Embodiments of the present invention relate to systems, methods and computer program products for evaluating customers' ability to manage revolving credit. To do so, for example, embodiments of the present invention review aggregated transaction data taken from a large number of revolving-credit accounts, and detect credit-usage patterns at an account level. The detected credit-usage patterns are then applied in risk models to evaluate individual customers' ability to manage revolving credit.