Machine Learning Identification of Order Data Across Merchant Webpages
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing rule-based approaches for identifying order-related data on diverse merchant websites are cumbersome and difficult to scale, requiring customization and frequent updates due to frequent layout changes, making them impractical for large numbers of merchants.
Innovation Solution
A machine-learning approach using neural networks to identify order-related data on webpages by converting HTML tags into vector representations, applying a trained model to predict labels, and outputting identified data, enabling scalability and robustness across various websites.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If a rules-based approach using regular expressions is used to identify order-related data, then it can work for individual merchant websites, but it becomes cumbersome and difficult to scale to thousands of merchants
Solution Approach 1:
The patent replaces the mechanical rules-based system with a machine learning model that automatically learns patterns from webpage data. Instead of manually creating and maintaining regular expressions for each merchant, the system uses trained neural networks to automatically identify order-related elements, eliminating the need for per-merchant customization while maintaining adaptability.
Solution Approach 2:
The machine learning model performs self-service by automatically adapting to different merchant websites without human intervention. The model learns from training data and can independently identify order-related elements on new merchant sites, eliminating the need for manual rule creation and updates for each merchant.
2Measurement precision
If custom rules are created for each merchant website, then order-related data can be accurately identified, but frequent layout changes require continuous monitoring and updates
Solution Approach 1:
The system performs preliminary action by pre-training the machine learning model on diverse webpage layouts before deployment. This allows the model to already understand various patterns and structures, so when deployed, it can accurately identify order-related data without requiring real-time rule updates when layouts change.
Solution Approach 2:
The patent introduces dynamics by making the system adaptable to changes through the machine learning model's ability to learn new patterns. Instead of static rules that break when layouts change, the dynamic ML model can adjust to new webpage structures, reducing the need for continuous monitoring and updates.
3Ease of manufacture
If traditional rule-based methods are used, then implementation is straightforward for a few hundred merchants, but scaling to thousands or tens of thousands of merchants becomes very difficult
Solution Approach 1:
The patent applies universality by creating a single machine learning model that can serve multiple merchants across different industries and webpage structures. This universal model replaces the need for separate rule sets for each merchant, enabling easy scaling from hundreds to thousands of merchants without proportionally increasing implementation complexity.
Data Source
AI summary
The present disclosure relates to a system, method, and computer program for identifying order-related data on a webpage using machine learning. The system identifies one or more HTML blocks on a webpage for further processing, where each HTML block includes one or more HTML tags that are each associated with metadata and a value. The system converts each of the HTML tags into a vector representation. The system applies a neural network model to each of the vector representations, resulting in another vector representation, which is converted into a machine-generated label prediction for each tag, where the neural network model is trained to predict labels, including order-related labels, corresponding to HTML tags. The system identifies order-related data on the webpage from the machine-generated label predictions for the HTML tags and the corresponding tag values. The system outputs the identified order-related data.