Identifying Comparative Chinese Cellphone Online Reviews Based On CR Emotional Lexicon Method

In the era of "wisdom +",the Internet has increasingly become the primary medium for us to understand the world,master information and obtain information.The emerging information technology has promoted the rise of smart retail.The traditional retail industry has undergone transformation and reform.The competition among online retailers is increasingly fierce,and the new territory needs to be rebuilt.On the other hand,with the increasing popularity of the Internet,especially the mobile Internet,the Internet is no longer just a source for users to obtain information,but also a channel for users to express opinions,exchange emotions and share information.In this context,with the rapid development of online retail,there have been a large number of online comments published by users,among which,comparative sentences occupy an increasingly important position in online comments.The comparative sentence in online comments refers to the sentence that expresses the difference in attributes or characteristics of different things(products).It is a common and persuasive way of expression in Chinese online comments.Through effective comparative sentence recognition method,useful knowledge can be mined to obtain the comparative advantages and disadvantages of products,and help consumers make better decisions and businesses to improve product quality.Therefore,the mining of comparative sentence information becomes an urgent research topic.However,there are still many missing points and shortcomings in the existing research.For example,the classification of comparative sentences is not fine,the accuracy of comparative sentence recognition is not high,and the research on cross-domain is insufficient.In order to solve these problems,this paper focuses on the characteristics of Chinese commentary,focusing on the scientific problem of "improving the recognition accuracy of comparative sentences and exploring cross-domain automatic identification".The systematic research is carried out from the perspectives of traditional classification model and CR(comparison result)sentiment lexicon,and therelated problems appearing in the research process are discussed in detail.The research in this paper includes the following two parts:(1)Chinese online commentary comparison sentence recognition based on traditional classification modelFeature selection plays an important role in comparative sentence recognition.The existing research mainly adopts the method of pattern recognition,which has certain effects,but requires a large number of manual rules.The third chapter introduces the text analysis method.With the help of the traditional classification model with better classification effect,this paper studies the combination of comparative elements in mobile online reviews,selects feature items,and finally realizes the identification of comparative sentences in online reviews,and discusses the influence of different types of comparative element combinations on the accuracy of comparative sentences.(2)Comparison sentence recognition of Chinese online comments based on CR emotion dictionaryIn order to make better use of the semantic information of vocabulary,further improve the accuracy of comparative sentence recognition,and explore cross-domain automatic recognition,the fourth chapter takes the most cross-domain comparison result words among the comparison elements as the seed words,and based on the synonyms and synonymy of words,constructs the CR emotion word database and determines the polarity value of emotion words.The CR sentiment lexicon is used to identify the sentence.The experimental results show that the constructed CR sentiment vocabulary performs well in the comparison of Chinese online comments,which improves the accuracy of comparative sentence recognition.