Research On Online Comment Driven Electronic Product Selection

The explosive growth in the number of online shopping sites and social media platforms has made it easy for consumers to share their reviews of products or services.Potential consumers use reviews to make better purchasing decisions for different products of the same type before buying.The in-depth research on the difficulty of sorting and selecting alternative products in the field of electronic products using online reviews is carried out.The main work of this paper are:(1)For the fine-grained sentiment analysis of review texts,use natural language processing tool Hanlp to preprocess the online reviews in this article.Based on the LDA topic model,mine the number of product attributes that consumers are interested in in the review text to determine the evaluation indicators.Frequently classifies attribute words into corresponding attribute word sets to establish attribute word lists.(2)Using the keywords in the attribute vocabulary as the core,the candidate sentiment words corresponding to the electronic product attribute words are extracted through the dependency syntax analysis technology,and the FastText model is used to train the word vector from the domain-related corpus,and the KNN algorithm is applied to the candidate sentiment words.Classification and screening were expanded to construct a domain sentiment dictionary suitable for the field of electronic products for sentiment analysis.(3)Aiming at the problem of determining the weight value of consumers in the optimal purchase of electronic products,ordinary consumers are unable to give accurate attribute index weight values.This article converts the text sentiment analysis results into the form of intuitive fuzzy numbers.Partially ordered sets have the advantage of being able to sort solutions without the need for weight values.The partial ordered set method based on improved intuitionistic fuzzy TOPSIS model is used to rank alternatives,and the Hasse diagram is used to visualize the ranking.The results obtained in this article expand the means and theoretical guidance of using online reviews to rank and select alternative electronic products.The proposed decision model has certain reference significance for the development of electronic product purchase decision aids,and further the research lay a certain foundation.This paper has 34 pictures,11 tables,and 102 related references.