| As the Internet and e-commerce have grown quickly,one of the hottest areas of sentiment analysis is the mining of online reviews of e-commerce platform users.The research objects cover a wide range of industries including those related to books,movies,hotels,tourism,electronic products,etc.Texts of food reviews should be more thoroughly researched.In addition,the main task of sentiment analysis at present includes the classification of positive and negative bipolar sentiment tendencies and less in-depth mining of the attitudinal resources embedded in the text at the semantic level.Based on this,this paper employs the attitude subsystem of appraisal theory as a theoretical framework,combined with sentiment analysis,to investigate the characteristics of attitude resource distribution and the variations in means of realization in Chinese and English online reviews of goji berries,as well as how reviewers express their attitudes and positions and establish connections with readers through attitude resources.In this study,1,000 Chinese and English online reviews of goji berries crawled from Taobao and Amazon shopping platforms serve as the corpus.The three steps of data collection and pre-processing,aspect extraction and sentiment analysis,and attitude resource annotation and analysis are used to study and analyze users’ sentiment tendencies toward health food goji berries as well as the distribution characteristics of attitude resources of evaluation discourse in online reviews.The three questions below are what this study aims to address:(1)What are the product aspects of goji berry that consumers pay attention to when shopping online? How satisfied are they with them?(2)What are the similarities and differences in the distribution characteristics and means of realization of attitudinal resources in Chinese and English goji berry online reviews?(3)How is the interpersonal meaning of online reviews realized through attitudinal resources?First,for e-commerce online reviews,this study uses Octopus collector to crawl Taobao and Amazon reviews,and Jieba and NLTK libraries for word segmentation,word frequency counting,and aspect extraction.Second,the SO-PMI algorithm is used to construct a sentiment dictionary in the goji berry domain,and sentiment tendency analysis is performed by Python to traverse sentiment words,degree adverbs,and negatives to calculate the sentiment value of each review containing aspect words and obtain consumer satisfaction with each product aspect.Finally,the attitude system of Martin’s appraisal theory is applied as the theoretical framework to classify,count and analyze attitude resources in reviews using the UAM Corpus Tool,and explore the attitude resources distribution and realization means in English-Chinese online reviews.(1)Through sentiment analysis,it is found that the five product aspects of Chinese and foreign consumers’ concern about goji berries are weighted from high to low as quality > packaging > price > logistics > service.And there are differences in the satisfaction of each aspect,with the overall satisfaction of English reviews ranging from high to low as quality > logistics > price > packaging > service,and the overall satisfaction of Chinese reviews ranging from high to low as quality > packaging >service > logistics > price.And through the experimental results,targeted suggestions are made to merchants to help them deepen their understanding of consumers’ shopping interests,to provide products that meet consumers’ needs.(2)In addition,relying on attitude system for the classification of Chinese and English evaluation discourse resources of goji berries,it is found that the overall distribution and means of realization of attitude resources are consistent,showing the distribution characteristics of more positive words than negative words,appreciation resources > affect resources >judgment resources.There are differences in the distribution and means of realization of attitude resources within the three subsystems.Reviewers tend to convey their opinions/attitudes toward products rather than inner emotions or judging people’s behaviors.(3)The lexical realization forms of interpersonal meaning in English and Chinese online reviews show the characteristics of adjective > verb > noun > adverb.Taking typical online reviews as an example,this paper analyzes the differences in the lexical choices of Chinese and English review writers in expressing sentimental attitudes and their influence on review readers in the context of business discourse,unearths emotional lexicon in the food domain within the framework of attitude system,and enriches the discourse study of online food reviews.This study offers an effective theoretical framework and empirical method for examining the similarities and differences in the distribution of Chinese and English online reviews attitudinal resource and means of realization by combining attitude system and big data mining techniques.Future business discourse analysis can be built on this theoretical framework and further extended according to the type of data to explore the characteristics and differences of evaluative discourse in different genres.This study has the limitation of only concentrating on the interpersonal meaning of the attitude system’s direct realization,future research could further investigate the implicit construction of evaluative discourses such as metaphor and semantic rhyme. |
PDF Full Text Request