Product Name Category Estimation Using Substring Graph Embeddings
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing methods for estimating product categories from product names are inaccurate due to the variability in substring selection and reliance on large language models not directly related to category estimation, leading to a burden in selecting categories for a large number of products.
Innovation Solution
An information processing device that extracts substrings from product names using category-representing strings, trains feature values, and estimates categories with higher accuracy by generating graphs and calculating embedding representations to minimize redundancy and leverage language models.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If machine learning techniques are used to extract feature values from product names and estimate categories, then the burden of manual category selection is reduced, but estimation accuracy deteriorates due to variability in substring selection and reliance on large language models not directly related to category estimation
Solution Approach 1:
The system performs preliminary extraction of substrings from product names before category estimation. By pre-processing the product names to extract meaningful substrings and building a graph structure in advance, the system prepares optimized input data for the estimation model, improving accuracy while maintaining automated operation.
Solution Approach 2:
The patent introduces an intermediary graph structure that connects substrings from product names to category information. This graph acts as a mediator between the raw product name data and the category estimation model, enabling more accurate category prediction by leveraging relationships between substrings and categories rather than relying solely on large language models.
2Extent of automation
If substrings are extracted from product names for category estimation, then automated category assignment becomes possible, but accuracy deteriorates due to variability in substring selection
Solution Approach 1:
The system performs preliminary extraction of substrings from product names before category estimation. By pre-processing the product names to extract meaningful substrings and building a graph structure in advance, the system prepares optimized input data for the estimation model, improving accuracy while maintaining automated operation.
Solution Approach 2:
The patent introduces an intermediary graph structure that connects substrings from product names to category information. This graph acts as a mediator between the raw product name data and the category estimation model, enabling more accurate category prediction by leveraging relationships between substrings and categories rather than relying solely on large language models.
Data Source
AI summary
An information processing device configured to: extract, for each of a plurality of pieces of first input information including a target string representing a target and a category string representing a category to which the target belongs, a first substring included in the target string, and extract, from the first substring, a second substring using the category string; determine a second substring associated with the target string included in the first input information; and train first feature information and second feature information to increase a matching rate between a category of the target string estimated from the second feature information and a category represented by the category string, the first feature information representing a feature of the determined second substring, the second feature information representing a feature of the target string including the determined second substrings, and being calculated based on the first feature information.