Study On The Evaluation System Based On The Emotion Of The Online Comments Of Scenic Spots

As China's Internet business model matures,residents' lives are becoming more convenient.Subsequently,it can be seen that the transaction scale of online market is also continuously increasing,and the social and commercial value of information comment is gradually increasing.How to mine the emotional information in the text and obtain the emotional tendency has become the focus of business intelligence.Through the emotional analysis of online comments,we can not only understand the degree of tourists' recognition of the scenic spot and provide Suggestions to the scenic spot,but also understand the needs of tourists,which is conducive to the development of business intelligence.This paper mainly completed the following work: first,online comment text information collection and preprocessing: using crawler technology to obtain colorful yunnan happy world online comment,and using word segmentation,part of speech tagging and dependency grammar analysis methods to extract sentence information,the use of syntactic structure to remove information redundancy.The second is the exploratory analysis of the collected information: in order to ensure that different dimensions of the same comment do not interfere with each other.This paper divides online comments into eight dimensions: traffic,ticket price,food and drink,entertainment items,queuing time,service attitude,landscape,shopping and accommodation.Meanwhile,it draws word cloud map to explore key high-frequency information.The third is to establish text emotion classification according to thesaurus: this paper adopts the method of combining text analysis with manual construction to establish the scenic spot emotion thesaurus,and divides the emotion words into five categories: super positive emotion,positive emotion,neutral emotion,negative emotion and super negative emotion.The evaluation dimension was classified and scored by matching the emotion words with the comment dimension.Fourth,the text emotion classification based on machine learning: the text classification algorithm is applied to train and classify the online comment emotion.First,each dimension of online comments was manually tagged with emotion,then the comments and tags were trained and classified to obtain the classification template,and the tested data were used to repeat the training 15 times through the machine learning method to obtain the classification results.With the development of artificial intelligence in recent years,machine learning has become an important method to establish emotion classification.In order to improve the accuracy of Chinese emotion classification,this paper adopts an emotion classification method which combines text information extraction and integration with classification algorithm,and improves the problem of inaccurate grasp of emotion and difficulty in multi-category classification in machine learning classification method.At the same time,this paper also carried out a case test comparison,and the test results showed that the accuracy rate of emotion classification based on manual selection and word bank establishment was higher than that of Chinese emotion classification based on machine learning.According to the review system,the advantages and disadvantages of the review dimension can be seen without specific review,which greatly saves the time cost of selection,avoids the travel minefield of tourists more efficiently,and points out the specific strategies and directions that businesses can improve.