Product Recommendation System Using Subjective Attribute Derivation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current product recommendation systems rely on objective data and fail to consider subjective attributes of customers and products, leading to irrelevant recommendations that may decrease sales by suggesting products customers do not need or want.
Innovation Solution
A system that derives subjective attributes for customers and products based on objective data, using a prompt-response process to narrow down recommendations, ensuring that only relevant products suited to the customer's preferences are presented.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If product recommendation systems use only objective data such as purchase history and browsing history, then the system can generate a large number of product recommendations, but the recommendations become irrelevant and may decrease sales
Solution Approach 1:
The patent transforms objective product attributes into subjective attributes by applying natural language processing and sentiment analysis to product reviews and descriptions. This changes the parameter type from quantifiable objective data to interpretive subjective characteristics, enabling the system to filter recommendations based on customer preferences rather than just purchase history.
Solution Approach 2:
The patent introduces subjective attributes as an intermediary layer between objective product data and recommendation generation. These subjective attributes act as a mediator that translates raw product specifications into preference-based filtering criteria, allowing the system to eliminate irrelevant products while maintaining a manageable recommendation list size.
2Productivity
If the system recommends products based on recent purchases, then it can suggest similar products quickly, but it may recommend products the customer already owns or does not need
Solution Approach 1:
The patent enriches the recommendation parameters by incorporating subjective attributes derived from product reviews and descriptions. Instead of relying solely on objective category matching, the system evaluates products based on subjective characteristics that reflect customer preferences, thereby improving recommendation accuracy without significantly impacting generation speed.
Solution Approach 2:
The patent implements a feedback mechanism where customer interactions with recommended products (purchases, views, skips) are used to refine the subjective attribute model. This continuous learning process improves the system's ability to distinguish between products a customer wants and those they already own or do not need, enhancing accuracy over time.
Data Source
AI summary
In some embodiments, apparatuses and methods are provided herein useful to recommending products to a customer based on derived subjective attributes for the products. In some embodiments, a system for recommending products to a customer comprises a customer prompt module configured to receive an indication of a category, determine, based on derived subjective attributes of the customer, an initial set of products, select a prompt based on an estimated number of products in the initial set of products that can be eliminated, incorporate, with a script, the prompt, present the prompt, receive the response to the prompt, a scoring module configured to calculate a score for the response to the prompt, a product recommendation module configured to eliminate at least a portion of the initial set of products, determine that a threshold has been reached, and a presentation module configured to generate a GUI including the remaining products.