Catalog Data Normalization via Unstructured Text Parsing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing electronic catalog systems face challenges in searching for products due to de-normalized data, where suppliers use different descriptions for the same products, leading to inefficient search results.
Innovation Solution
A system that uses machine learning algorithms to identify and normalize product attributes from unstructured text, creating a standardized format for catalog data, which is then indexed for efficient search and retrieval.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If suppliers use supplier-specific product descriptions in electronic catalogs, then each supplier can maintain their own product terminology and branding, but the search engine fails to identify matching products across different suppliers due to de-normalized data
Solution Approach 1:
The patent introduces a normalization layer as an intermediary between supplier-specific product descriptions and the search engine. This normalization layer converts diverse supplier descriptions into a standardized format using mapping rules and synonym dictionaries, allowing the search engine to reliably match products across suppliers while preserving the original supplier-specific descriptions in the catalog data
Solution Approach 2:
The patent transforms the search process by changing the parameter format from raw supplier-specific descriptions to normalized standardized terms. By applying normalization rules that convert various product terminology into a common vocabulary, the system enables consistent search matching while maintaining the ability to retrieve results from multiple suppliers
2Adaptability or versatility
If product descriptions are stored in unstructured format from multiple suppliers, then the catalog can accommodate diverse product lines and supplier variations, but the data becomes de-normalized making search and retrieval inefficient
Solution Approach 1:
The patent segments the product description data into structured fields (product name, description, attributes) and applies normalization to each segment. This segmentation allows the system to maintain diverse supplier inputs while organizing the data into a standardized structure that improves search efficiency without losing catalog diversity
Solution Approach 2:
The normalization layer acts as an intermediary that processes unstructured supplier descriptions into structured normalized data. This intermediary layer enables the system to accommodate diverse product lines and supplier variations while simultaneously improving search efficiency through standardized data organization
3Device complexity
If the search engine queries product descriptions directly without normalization, then the system maintains simplicity in the search process, but it cannot match user input with product descriptions from different suppliers using different terminology
Solution Approach 1:
The patent introduces a normalization intermediary between the search engine and product descriptions. This intermediary automatically converts both user search queries and product descriptions into normalized terms, maintaining search process simplicity from the user perspective while dramatically improving product matching accuracy across suppliers through standardized term comparison
Data Source
AI summary
Provided is a method and system for normalizing catalog item data to create higher quality search results. In one example, the method may include receiving a record comprising an unstructured description of an object, identifying a type of the object from among a plurality of object types and identifying a predefined attribute of the identified type of object, extracting a value from the unstructured description corresponding to the predefined attribute and modifying the extracted value to generate a normalized attribute value, and storing a structured record of the object in a structured format comprising a plurality of values of a plurality of attributes of the object from the unstructured description including the normalized attribute value for the predefined attribute of the object.