Review Keyword Extraction Using Language Models and Spam Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Online shoppers face challenges in finding reliable product reviews amidst exaggerated or promotional content, leading to increased time and effort in searching for relevant information, and uncertainty in purchase decisions due to unreliable review data.
Innovation Solution
A language model-based method and system for extracting product review keywords by collecting review data, generating response data from predetermined questions, and filtering out spam/promotional content to enhance the reliability and accuracy of review keywords.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If all product reviews are collected and analyzed, then the quantity of review data increases, but the reliability of extracted keywords decreases due to inclusion of spam and promotional content
Solution Approach 1:
The patent extracts and removes spam and promotional review data from the collected review dataset using a classification model. This selective extraction eliminates harmful content while preserving legitimate reviews, thereby maintaining high keyword reliability even when processing large volumes of review data.
Solution Approach 2:
The patent introduces a classification model as an intermediary between review data collection and keyword extraction. This intermediary component filters and categorizes reviews to identify and remove spam/promotional content, enabling reliable keyword extraction from large-scale review datasets without being contaminated by false information.
2Reliability
If manual review analysis is performed to ensure reliability, then the reliability of review keywords improves, but the time and effort required increases significantly
Solution Approach 1:
The patent replaces manual mechanical review analysis with an automated language model system. The model automatically processes review data, generates responses to predetermined questions, and extracts keywords without human intervention, achieving both high reliability and efficiency through automated intelligent processing rather than manual examination.
Solution Approach 2:
The patent enables the review analysis system to serve itself by automatically filtering spam, generating meaningful responses, and extracting keywords without requiring manual verification. The classification model and language model work autonomously to ensure reliability while minimizing time and effort investment from users.
3Productivity
If a language model generates responses from review data, then the extraction efficiency improves, but the complexity of the processing system increases
Solution Approach 1:
The patent segments the review processing system into distinct functional modules: a classification model for spam detection, a language model for response generation, and a keyword extraction component. This segmentation allows each module to perform its specific function efficiently while maintaining overall system manageability and reduced complexity through modular architecture.
Data Source
AI summary
A language model-based method for extracting a product review keyword includes collecting, on the basis of information for specifying a product, review data associated with the product; using a language model so as to generate, on the basis of a plurality of predetermined questions, at least one piece of response data from at least some pieces of the review data; and extracting, on the basis of the at least one piece of response data, a review keyword associated with the product.