Research on Product Attribute Extraction and Classification Method for Online Review

Abstract

With the rapid development of e-commerce, a wide variety of product reviews have appeared on the internet. These reviews not only provide consumers with a reference, but also help manufacturers make reasonable marketing decisions. In online reviews, customers usually give opinions on multiple attributes of products, therefore, the analysis of product attributes is a crucial issue in product review analysis. This paper studies in depth the extraction and classification of product attribute words from the context. Aiming at the colloquial speech in the online review and the incompleteness of the existing dictionary-based word segmentation methods, this paper uses machine learning method to identify product attribute words. By introducing the word internal tag method to identify the segmented out-of-vocabulary attribute words, and add it to the user's dictionary, correcting the word segmentation results. In addition, a word-level text classification method based on distributed word representation is proposed, and the semantic and syntactic features of the word vectors are used to classify the product attribute words.