Research On Online Review Quality Classification Based On Am?BiLSTM

With the popularity of the Internet and the rapid development of e-commerce,ecommerce has been integrated into people's daily life by virtue of its low price,convenient transaction,and other advantages.However,this shopping mode has also brought many problems,among which the most obvious problem is that consumers cannot experience and study the quality and performance of goods as intuitively and truly as physical shopping.Therefore,consumers increasingly rely on online reviews to make shopping decisions,and the importance of online reviews is becoming increasingly prominent.With the appearance of business praise and malicious bad reviews,it is difficult for many consumers to identify the useful high-quality reviews from many low-quality reviews.In this case,identifying high-quality reviews to help consumers choose,and identifying lowquality reviews that are not helpful to consumers become a new direction of online review content research.At present,some scholars have begun to study the quality evaluation of online reviews,and how to use various methods to achieve the quality classification of online reviews.However,most of the online review quality classification algorithms are based on machine learning,and its main content is to determine the various characteristics that determine the quality of online reviews.Although the quality classification effect of this method is acceptable,it takes a lot of time and energy to consider the feature selection and extraction method,and ignores the importance of the characteristics of online comment text in determining the quality of online comments.With the rapid development of deep learning,the research of online review quality classification has a new direction.According to the characteristics of online reviews,this paper makes full use of the text features of online reviews.Firstly,according to the text features of online reviews,an online review quality evaluation index system is established;Then,considering the importance of product attribute words for online comment quality classification,a domain dictionary is constructed to pre label online comments,which can assist the process of product attribute words extraction and improve the accuracy of product attribute words extraction.Then the product attribute words are integrated into the feature extraction process by using Bi LSTM model and attention mechanism,which improves the importance of product attribute words in the process of quality classification and the accuracy of online comment quality classification.At the end of this paper,we select the online comment data set crawled from Jingdong Mall as an example.First,we use the online comment quality evaluation index system to classify the quality of the online comment set,and then use the domain dictionary to extract the product attribute words.Finally,the model is used to classify the quality of online reviews,and the results of the model are greatly improved compared with the baseline model.The effectiveness of the proposed online review quality classification model is verified.At the same time,this paper provides reference application strategies for businesses and platforms from online reviews of different quality categories.