Research On Emotional Analysis Of Fine-grained Attributes Of Online Reviews

The emotional tendency to obtain online reviews is one of the most important research areas in data mining.However,coarse-grained emotional classification algorithms are difficult to excavate the emotional tendencies of fine-grained attributes in complex online reviews.How to accurately excavate emotional tendencies in online reviews is a difficult problem in data mining.Starting from the problem of accurately obtaining the sentiment tendency of reviews,this paper proposes a model to get the fine grained attributes and the polarity analysis of the goods,and studies how to get the fine grained emotional tendencies of online reviews.This paper first proposes a web crawler for online reviews,and takes the car's home website as an example to crawl 814025 online reviews.Then,in order to get the fine-grained attributes of products in online reviews,a semi supervised learning method for fine-grained attribute acquisition is proposed,and the fine grained attribute set of automobiles is acquired by taking the crawling online reviews of large automobiles as an example.Finally,a sentiment classification method based on CRF and SVM is proposed.Experiments are carried out on crawling car online reviews to get fine-grained attributes and emotional tendencies in online reviews.In this paper,a semi supervised learning based fine grained attribute extraction model is proposed.Compared with traditional methods based on seed words or LDA,the paper method can get more meaningful fine grained attributes.Compared with the latest literature methods,the paper uses a semi supervised method to improve the efficiency of fine-grained extraction.Second,the proposed sentiment analysis method integrating CRF and SVM has an average accuracy of 90%.Compared with the classification without extracting feature fragments,the algorithm proposed in this paper is better.Third,the combination of fine-grained extraction and sentiment classification method,presents a method of online reviews of fine-grained sentiment classification,get online comments are fine-grained attributes of emotion tendency,improve the online reviews existing research in rough classification emotional problems.The research results of this paper can extract more detailed emotional tendencies from online reviews and help consumers or enterprises make more effective decisions,which has important theoretical and practical significance.