Text-Based Product Matching Using NLP for Retail Pricing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional methods for competitive pricing in the retail industry rely on image similarity approaches, which are costly, require significant computational power, and are not scalable for new products or product variants, leading to inaccurate matching results.
Innovation Solution
A processor-implemented method and system for text-based target product matching using natural language processing to generate product corpora, identify unique context tokens, calculate cogent scores, and build a similarity matrix for accurate product matching without the need for image data or extensive labeling.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If image similarity or image matching based approaches are used for comparing competitor products pricing, then product matching can be performed, but significant computational power and infrastructure costs are required
Solution Approach 1:
The patent replaces image-based mechanical vision systems with text-based natural language processing systems. Instead of using image similarity algorithms that require significant computational power, the invention uses NLP techniques to extract and compare product attributes from text descriptions, thereby reducing computational resource requirements while maintaining matching accuracy.
Solution Approach 2:
The patent changes the data representation parameter from image data to text data. By transforming product information into structured text formats and using NLP to process this text, the system achieves efficient computation without the heavy computational burden of image processing, while still enabling accurate product matching across different retailers.
2Adaptability or versatility
If image similarity approaches are used for product matching, then existing infrastructure can be utilized, but the system is not scalable for new products and enriched data information
Solution Approach 1:
The patent creates a universal text-based product representation system that can handle diverse product types and categories through a single NLP-based framework. The system uses a general-purpose text processing architecture that automatically adapts to new products without requiring reconfiguration of the underlying infrastructure, enabling seamless scalability across different product domains.
Solution Approach 2:
The patent uses text descriptions as a simplified copy or representation of product information, replacing the need for complex image processing infrastructure. This text-based copying approach allows the system to easily accommodate new products by simply inputting their text descriptions, without requiring additional computational infrastructure or complex configurations.
3Measurement precision
If conventional ML or DL based models are used for competitor product matching, then product matching can be performed, but large amount of labelling data and frequent training are required
Solution Approach 1:
The patent implements a self-service text processing system that automatically extracts and compares product attributes from unstructured text descriptions without requiring manual labelling. The NLP system autonomously processes product information, identifies relevant features, and performs matching tasks independently, eliminating the need for large labelled datasets and reducing the requirement for frequent model retraining.
4Measurement precision
If image-based approaches are used for product matching, then visual similarity can be captured, but extracting key attributes from images is challenging especially with low quality images
Solution Approach 1:
The patent substitutes image-based attribute extraction with text-based extraction using NLP techniques. Instead of attempting to extract attributes from potentially low-quality images, the system processes text descriptions that directly contain product attribute information, thereby achieving accurate attribute extraction independent of image quality and eliminating the harmful dependency on image quality.
Data Source
AI summary
This disclosure relates generally to methods and systems for text-based target products matching to keep competitive price for increasing customer traffic. The conventional image similarity or image matching based approaches require appropriate configurations and infrastructure, quality images and more computational power. Also, the configurations and infrastructure may not be scalable for new products and may not be suitable for all retail industry types. The present disclosure uses only the text-based data inputs of the retailer and the competitor. The present disclosure first generates a data corpus from the data inputs of the retailer and the competitor. Then, context tokens are identified using a natural language-based technique and a cogent matrix is formed based on the data inputs of the retailer and the competitor, the context tokens, and using a cogent score calculation function, to get the matching results.