Reinforcement Learning for Dynamic Product Combination and Discount Rate Optimization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Online purchases face challenges in recommending products with suitable discount rates to users, especially considering low purchase probability and repurchase likelihood, as existing methods fail to effectively leverage user feedback and behavior data in real-time promotions.
Innovation Solution
An electronic device employs reinforcement learning to identify user purchase intentions and infer product combinations with discount rates by processing user and product data through AI models, providing personalized promotions based on user behavior and product relationships.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If online services provide promotions to users, then user engagement may be improved, but the low purchase probability and low repurchase probability make it difficult to achieve effective sales conversion
Solution Approach 1:
The patent implements a feedback mechanism where the AI model continuously learns from user responses to promotions. User feedback (purchases, views, ignores) is fed back into the reinforcement learning model to update purchase intention predictions and optimize future promotion strategies, creating a closed-loop system that improves conversion effectiveness over time
Solution Approach 2:
The system dynamically adjusts promotion parameters (discount rates, product combinations, timing) based on AI model predictions of user purchase intention. By changing these parameters according to predicted user behavior, the system optimizes the balance between engagement and actual sales conversion
2Productivity
If generic promotions are provided to all users, then implementation complexity is reduced, but user engagement and purchase effectiveness decrease
Solution Approach 1:
The AI model automatically performs data collection, analysis, and promotion optimization without manual intervention. The reinforcement learning system self-adjusts promotion strategies based on learned patterns from user behavior data, reducing the need for complex manual personalization while maintaining high engagement levels
Solution Approach 2:
The patent employs a universal AI model that handles multiple functions: collecting user data, predicting purchase intention, optimizing product combinations, and determining discount rates. This multi-functional approach consolidates complexity into a single system rather than requiring separate mechanisms for each personalization aspect
3Reliability
If discount rates are increased to improve purchase probability, then sales conversion may improve, but profit margins decrease
Solution Approach 1:
The system applies different discount rates to different user-product combinations based on predicted purchase intention. High-intention users receive lower discounts while maintaining high conversion probability, whereas low-intention users receive higher discounts to stimulate interest, optimizing the balance between purchase probability and profit margin for each specific case
Data Source
AI summary
A method by which an electronic device provides a service to a user, includes: obtaining data related to at least one of the user, a plurality of products, or one or more marketing activities; identifying the user's purchase intention based on the data; identifying at least one product combination comprising two or more products from among the plurality of products and a discount rate of the at least one product combination by applying the identified user's purchase intention and the data to an artificial intelligence (AI) model; and displaying, on a display of the electronic device, the at least one product combination and the discount rate.