A Study On Satisfaction Of Real-life Escape Room Games Based On Online Review

On the premise that the basic material needs of life are met,the public’s demand for spiritual life is becoming stronger and stronger.Traditional entertainment methods such as watching movies and KTV singing have already caused people to have aesthetic fatigue,and the novelty of real person secret room escape can meet people’s demand for stimulation at the spiritual level The market scale of the secret room escape industry has reached 10.12 billion in 2021,with 2.8 million consumer trips and more than 10000 stores.It occupies an important position in entertainment activities and enriches the entertainment life of the public Nowadays,with the rapid development of the Internet,a large number of consumers buy tickets online and comment.Therefore,this paper selects the online comment text of secret room escape as the research object to mine valuable information Through the research and analysis of the secret room escape industry,it can not only help consumers choose high-quality merchants,but also help merchants understand consumers’needs more deeply,and improve stores in order to improve consumers’ satisfaction.Therefore,the research on the above problems is of great significance to consumers and merchants.Based on the online comment data of human escape from the secret room from January 2021 to January 2022,this paper studies the emotional classification of comment texts and consumer satisfaction In the research of emotion classification of comment text,firstly,preprocess the original data,including deleting useless information,classifying positive and negative emotions of each comment text,and divide some of the preprocessed data into training set and test set Then,on the training set,LSTM model,Bert model,Bert-LSTM model and Bert-BiLSTM model are used for emotion classification task.The analysis results show that Bert-BiLSTM model has the strongest generalization ability.Therefore,this model is used as the optimal emotion classifier,and this model is used to classify the positive and negative emotions of other text data In the study of consumer satisfaction,firstly,LDA theme mining is carried out for positive and negative comments respectively,and the characteristics concerned by consumers are analyzed.These characteristics are classified into five attributes to obtain the basic attribute feature table,of which the five attributes are site layout,service,experience,script and cost performance On this basis,by using the vocabulary similarity calculation of HowNet and expanding the basic attribute thesaurus based on the extended version of synonym forest of Harbin University of technology,a characteristic dictionary only for the secret room escape industry is obtained,which is conducive to the supplement and improvement of the secret room escape industry according to the needs of consumers Finally,the emotional scores of each store in five attribute characteristics are calculated,the satisfaction of consumers to each store is analyzed,and the advantages and disadvantages of stores in each attribute are studied.The results show that the factors affecting consumer satisfaction mainly include the characteristics of five attributes,and these characteristics can be used as the standard to judge the advantages and disadvantages of businesses,which can not only help consumers quickly select excellent businesses,but also help businesses to make targeted improvement,form benign competition among businesses and promote the high-quality development of the industry.