E-commerce Pricing Extraction via Visual Map Stylistic Heuristics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for determining pricing and availability information for e-commerce entities saved at third-party wish lists are inaccurate due to inconsistent formatting across different websites and inability to track multiple websites, often providing incorrect information to users.
Innovation Solution
An application that uses stylistic heuristics in addition to text analysis to identify and update pricing and availability information by generating a visual map representation of e-commerce web pages, applying stylistic identifiers, and using machine learning or rules-based models to determine the likelihood of candidate elements being pricing or availability elements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If conventional text-only analysis methods are used to determine pricing and availability information, then the system is simple to implement, but the accuracy of pricing and availability information is poor
Solution Approach 1:
The patent transitions from one-dimensional text analysis to two-dimensional visual map analysis by incorporating spatial position, font size, color, and other visual attributes of elements on the web page. This dimensional expansion enables the system to distinguish pricing and availability information more accurately by considering both content and visual presentation characteristics.
Solution Approach 2:
The patent introduces a visual map representation as an intermediary between the raw web page and the pricing/availability information extraction. This visual map serves as a mediator that captures both the textual content and visual characteristics of elements, enabling more accurate identification of pricing and availability information while maintaining system modularity.
2Adaptability or versatility
If only text content is analyzed to identify pricing information, then the processing is fast and simple, but the system cannot distinguish between pricing information in different formats from different websites
Solution Approach 1:
The patent creates a universal visual map representation framework that can handle multiple website formats and structures. By extracting and standardizing visual attributes (position, font size, color) across different websites, the system achieves format-agnostic pricing and availability information extraction, making it versatile across diverse e-commerce platforms.
Solution Approach 2:
The patent changes the parameters used for information extraction from purely textual parameters to include visual parameters such as element position, font size, color, and hierarchy. This parameter expansion enables the system to adapt to different website formats by recognizing pricing and availability information through its visual characteristics rather than relying on consistent text patterns.
3Quantity of substance
If the system tracks entities from multiple different websites, then the coverage of entities is comprehensive, but the difficulty of accurately tracking and updating pricing information increases
Solution Approach 1:
The patent segments the web page into a structured visual map with distinct elements representing different information types. By dividing the complex web page into manageable visual components (headers, prices, availability indicators), the system can accurately track and update pricing information for multiple entities across different websites while maintaining precision through consistent visual attribute analysis.
Data Source
AI summary
Disclosed is a system and method for determining pricing and availability of e-commerce entities. The determination includes accessing a web page that includes an entity of interest and generating a visual map representation of the web page that includes elements related to the entity's pricing/availability and stylistic identifiers that are applied to those elements. Stylistic identifiers are pricing/availability identifiers based on stylistic heuristics. Certain elements are identified as candidate elements and a likelihood that a candidate element is pricing/availability element of the entity is determined for the candidate elements. A candidate element is selected as the pricing/availability element of the entity by analyzing the determined likelihood of various candidate elements.