Hierarchical Transaction Classification for Diverse Data Formats
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional data enhancement techniques struggle to handle diverse and cryptic transaction records generated by different devices, leading to challenges in data mining and analysis, especially in the financial technology industry, due to the lack of a universal standard and frequent variations in transaction formats.
Innovation Solution
A hierarchical classification system that identifies tag sequences from transaction records, clusters them, and uses cluster-level and cluster-specific classifiers to predict labels, enabling flexible and accurate analysis of transaction data even when formats differ.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If conventional data enhancement techniques are used to process transaction records, then processing simplicity is maintained, but the ability to handle diverse data formats from different devices deteriorates
Solution Approach 1:
The patent segments the classification process into multiple hierarchical levels: first identifying transaction type, then merchant category, then specific merchant. This multi-stage segmentation allows the system to handle diverse data formats by breaking down the complex classification task into manageable steps, where each stage handles specific aspects of the transaction record.
Solution Approach 2:
The patent introduces a hierarchical dimension to the classification process, moving from flat single-level classification to multi-level hierarchical classification. This dimensional change enables the system to accommodate diverse data formats by organizing classification criteria in a structured hierarchy that can adapt to variations in input data.
2Quantity of substance
If transaction records are stored in cryptic shortened format to save space, then storage efficiency is improved, but data mining accuracy deteriorates
Solution Approach 1:
The patent performs preliminary classification and enrichment of transaction records during the data ingestion phase. By pre-processing and categorizing transactions into standardized hierarchical categories before storage, the system maintains storage efficiency while ensuring that the essential classification information is captured and preserved for accurate later analysis.
Solution Approach 2:
The patent introduces standardized merchant category codes and transaction type codes as intermediary representations between the original cryptic transaction data and the analysis systems. These standardized codes serve as a mediator that preserves storage efficiency while enabling accurate data mining by providing consistent, interpretable categories.
3Adaptability or versatility
If service providers frequently change coding strategies and partnerships, then business adaptability is improved, but system stability deteriorates
Solution Approach 1:
The patent implements a universal hierarchical classification framework that can accommodate multiple coding strategies and data formats from different service providers. The standardized category hierarchy serves as a universal interface that remains stable even as underlying data sources change, allowing the system to handle various partnerships and coding schemes without requiring fundamental changes to the classification infrastructure.
Data Source
AI summary
Methods, systems and computer program products implementing hierarchical classification techniques are disclosed. A hierarchical classification system receives training data including labeled transaction records. The system determines tag sequences from the training data. The system clusters the tag sequences into clusters. The system determines a cluster-level classifier that is trained to predict a cluster for an input transaction record. The system determines a respective cluster-specific classifier for each cluster. The system trains the cluster-specific classifier to predict a label of entity of interest for an input transaction record, given a particular cluster. Upon receiving a test transaction record, the system first applies the cluster-level classifier to determine a particular cluster for the test transaction record, and then determines a label of entity of interest of the test transaction record by applying a cluster-specific classifier of that particular cluster.