Data Augmentation Matching Using Seed Embeddings Without Common IDs
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Entities face challenges in identifying demographic trends for customers and prospective customers due to limited matching between their customer data sets and third-party survey data sets, which often lack a common identifier, resulting in insufficient demographic and behavior data for tailored products and services.
Innovation Solution
A computer-implemented method and system for data augmentation that uses a data augmentation model and seed entries to match and augment candidate data items with additional information from third-party data sets, incorporating neural networks and embedding vectors to enhance user data with demographic and behavior data.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If entities use traditional matching methods based on common identifiers to link customer data with third-party survey data, then matching accuracy is improved, but the quantity of matchable data deteriorates due to lack of common identifiers
Solution Approach 1:
The patent introduces hash values of user features (such as device identifiers, application identifiers, or user profile attributes) as intermediary keys to bridge the gap between customer data and survey data. Instead of requiring direct common identifiers, the system hashes relevant features from both data sources and matches on these derived hash values, enabling linking without direct identifier overlap
Solution Approach 2:
The system transforms user features into hash values, changing the parameter representation from raw identifiers to hashed forms. This parameter transformation allows matching to occur on derived representations rather than original identifiers, expanding the pool of matchable records while maintaining matching precision through the deterministic nature of hashing
2Loss of information
If entities augment data with multiple features to improve demographic insights, then the quality of demographic trends identification is improved, but the complexity of data processing deteriorates
Solution Approach 1:
The system performs preliminary actions by pre-computing and storing hash values of user features in both customer data and survey data before the actual matching process. This advance preparation eliminates the need for complex real-time feature extraction and comparison during matching, reducing processing complexity while enabling multi-feature augmentation
Solution Approach 2:
The patent segments the data processing into distinct stages: feature extraction, hash computation, matching based on hash values, and data augmentation. This segmentation allows each stage to be optimized independently, managing complexity through modularization while supporting comprehensive multi-feature data augmentation
Data Source
AI summary
Systems, methods, and computer readable media for data augmentation are described. The system comprises a network device, a memory comprising a data augmentation model and a plurality of seed entries, and a processor in communication with the network device and the memory. The processor is configured to receive a candidate data item in a second data set, generate a candidate seed corresponding to the candidate data item, and determine a data feature, based on the data augmentation model, for the candidate seed. Additionally, the processor is configured to generate at least one matching seed in the plurality of seed entries, the at least one matching seed based on the data feature. The processor is further configured to augment the candidate data item with data corresponding to the at least one matching seed.