3D Preference Visualization for E-commerce Product Clustering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
E-commerce websites struggle to provide personalized and effective data representation for users, as existing methods fail to organize product information in a way that allows users to easily find the best match based on their preferences, leading to a cumbersome shopping experience due to overwhelming and irrelevant data presentation.
Innovation Solution
A framework that acquires user preference information, generates rank scores for objects, groups them into clusters, and creates a three-dimensional visualization with optimized features such as size, color, and transparency to highlight relevant objects, allowing users to quickly identify strengths and weaknesses.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Loss of information
If traditional table format is used to display product information, then all product attributes can be shown, but users cannot easily find the best match based on their preferences
Solution Approach 1:
The patent transforms the traditional two-dimensional table display into a three-dimensional visualization space where products are represented as points or objects in 3D space. The three dimensions correspond to different preference criteria, allowing users to visually navigate and filter products based on multiple preferences simultaneously. This dimensional transformation enables users to quickly identify the best match by visual inspection rather than scanning through tabular data.
Solution Approach 2:
The system dynamically adjusts the visualization based on user preferences. As users modify their preference criteria, the 3D visualization automatically updates to reflect the new preference landscape, repositioning products in the visualization space according to their updated rankings. This dynamic adaptation allows the display to remain optimized for user needs without requiring manual reconfiguration.
2Loss of information
If in-depth tables are used to compare each attribute, then detailed product information is available, but the display becomes overwhelming and irrelevant attributes are shown
Solution Approach 1:
The patent applies local quality by making different regions or aspects of the visualization have different properties. The 3D space is organized such that different regions represent different preference criteria, and products are positioned according to their local characteristics along each dimension. This allows users to focus on specific attributes by examining relevant regions of the visualization without being overwhelmed by all attributes simultaneously.
Solution Approach 2:
The system segments the product information display into multiple dimensional layers or planes, where each dimension represents a specific preference criterion. Instead of presenting all attributes in a single overwhelming table, the information is segmented across three dimensions, allowing users to progressively explore products based on their priority preferences and drill down into specific attributes only when relevant.
3Adaptability or versatility
If multiple sorting criteria are allowed, then user preferences can be accommodated, but the global view of objects becomes difficult to maintain
Solution Approach 1:
By mapping multiple sorting criteria onto three spatial dimensions, the system maintains a coherent global view of all products simultaneously. Each dimension represents a preference criterion, and products are positioned in 3D space according to their performance across all criteria. This spatial representation allows users to grasp the overall product landscape and relative positioning of products across multiple dimensions at once, rather than switching between separate sorted lists.
Data Source
AI summary
Described herein is a technology for facilitating preference-based data representation. In accordance with one aspect of the technology, preference information is acquired from a user. Rank scores of objects are generated based at least in part on the user preference information. The objects are grouped into one or more clusters of objects based on the rank scores. A visualization of the one or more clusters of objects is then generated.