Affinity Profile Modeling From Transaction Embeddings
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Customer surveys for analyzing shopping trends are time-consuming, expensive, and prone to sampling errors, making them inefficient for inventory management and marketing efforts.
Innovation Solution
A system using machine-learning models to generate affinity profiles based on transaction data, extracting user and merchant embeddings without personal information, and determining similarities between them to predict purchasing behaviors and preferences.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If customer surveys are used to analyze shopping trends, then marketing insights can be obtained, but the process is time-consuming and expensive
Solution Approach 1:
The patent replaces manual survey administration with an automated machine learning system that processes transaction data to generate affinity profiles. The system uses neural networks to automatically analyze purchasing patterns, eliminating the need for time-consuming survey processes while maintaining comprehensive shopping trend analysis capabilities.
Solution Approach 2:
The system enables self-service analysis by automatically processing existing transaction data to generate customer affinity profiles without requiring active participation from customers. The machine learning model autonomously extracts insights from purchase history, replacing the need for customers to manually complete surveys.
2Loss of information
If customer surveys are used to analyze shopping trends, then marketing insights can be obtained, but the cost increases
Solution Approach 1:
The patent utilizes already-existing transaction data that would otherwise be underutilized resources. Instead of investing in expensive survey infrastructure, the system processes existing purchase records through machine learning models to generate affinity profiles, effectively using free or low-cost data sources to achieve marketing insights.
Solution Approach 2:
The system replaces expensive survey administration and analysis infrastructure with automated machine learning processing of transaction data. The neural network models process existing data efficiently, eliminating the need for costly survey platforms, sampling mechanisms, and manual analysis processes.
3Loss of information
If customer surveys are used to analyze shopping trends, then marketing insights can be obtained, but sampling errors occur
Solution Approach 1:
The system performs self-service analysis by automatically processing complete transaction records to generate affinity profiles. The machine learning model analyzes actual purchasing behavior data without requiring sampling, eliminating sampling errors while providing comprehensive and reliable insights into customer preferences and shopping patterns.
Solution Approach 2:
The patent replaces survey-based sampling mechanisms with machine learning processing of complete transaction datasets. The neural network models analyze actual purchase behavior data to generate affinity profiles, eliminating the sampling errors inherent in survey-based approaches while maintaining comprehensive coverage of customer behavior.
4Productivity
If machine-learning models are used to generate affinity profiles, then analysis efficiency improves, but model complexity increases
Solution Approach 1:
The patent segments the affinity profile generation process into distinct components: transaction data processing, neural network modeling, affinity profile generation, and similarity calculation. This modular approach manages model complexity by breaking down the overall system into manageable segments that can be developed and optimized independently.
Solution Approach 2:
The system introduces transaction data as an intermediary that bridges customer behavior patterns and affinity profiles. The machine learning models process this intermediary data to generate simplified affinity representations, managing complexity by using structured data as a mediator between raw transaction records and final marketing insights.
Data Source
AI summary
A system comprising one or more processors and a computer-readable, non-transitory medium including instructions which, when executed by the one or more processors, cause at least one of the one or more processors to obtain transaction data including a first merchant and a first user, and train a machine-learning model using the transaction data by executing the machine-learning model using as input the transaction data to generate a first merchant embedding corresponding to the first merchant and a first user embedding corresponding to the first user determining a similarity between the first merchant embedding and the first user embedding, and updating the machine-learning model based on whether the first user made a purchase at the first merchant within a predetermined time period.