LANGUAGE MODEL-BASED METHOD AND SYSTEM FOR EXTRACTING PRODUCT REVIEW KEYWORD
The present disclosure provides a language model-based method by which at least one processor extracts a product review keyword. The method comprises the steps of: collecting, on the basis of information for specifying a product, review data associated with the product; using a language model so as...
Saved in:
Main Authors | , , , , , , , |
---|---|
Format | Patent |
Language | English French Korean |
Published |
12.09.2024
|
Subjects | |
Online Access | Get full text |
Cover
Loading…
Summary: | The present disclosure provides a language model-based method by which at least one processor extracts a product review keyword. The method comprises the steps of: 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.
La présente divulgation concerne un procédé basé sur un modèle de langage par lequel au moins un processeur extrait un mot-clé d'évaluation de produit. Le procédé comprend les étapes consistant à : collecter, sur la base d'informations pour spécifier un produit, des données d'évaluation associées au produit ; utiliser un modèle de langage de façon à générer, sur la base d'une pluralité de questions prédéterminées, au moins un élément de données de réponse à partir d'au moins certains éléments des données d'évaluation ; et extraire, sur la base du ou des éléments de données de réponse, un mot-clé d'évaluation associé au produit.
본 개시는 적어도 하나의 프로세서에 의해 수행되는, 언어 모델 기반 상품 리뷰 키워드 추출 방법을 제공한다. 이 방법은, 상품을 특정하기 위한 정보에 기초하여, 상품과 연관된 리뷰 데이터를 수집하는 단계, 언어 모델을 이용하여, 미리 결정된 복수의 질문에 기초하여, 리뷰 데이터 중 적어도 일부로부터 적어도 하나의 답변 데이터를 생성하는 단계 및 적어도 하나의 답변 데이터에 기초하여, 상품과 연관된 리뷰 키워드를 추출하는 단계를 포함한다. |
---|---|
Bibliography: | Application Number: WO2024KR02603 |