An Empirical Study On Individual Risk Perception Factors And Their Impact On Leisure Consumption Under The COVID-19

The COVID-19 epidemic is a public health emergency with long duration and wide impact,which has a huge impact on national life.The current research on individual risk perception,emotion and behavior of major emergencies mainly focuses on the relationship between them.The interaction mechanism between the three is not clear,and the evaluation system is not perfect.This paper explores the relationship among individual risk perception,leisure consumption emotion and leisure consumption behavior,clarifies the interaction mechanism among the three,provides reference for relevant government departments to formulate policies,and provides suggestions for the rapid recovery of leisure consumption industry.Based on the TPB theory,this paper adds three dimensions of perceived positive utility,risk perception and individual factors,and develops the ’ Individual Leisure Consumption Behavior Scale in the Normalized Management of the New Coronavirus Epidemic ’.Combined with the scale,a questionnaire on leisure consumption behavior and emotion was designed.Through the multi-level sampling method,a sample survey was conducted on 60 community residents in the four major administrative regions of Hefei to obtain the survey data of individual leisure consumption behavior and emotion from August to October in 2022.Based on the survey data,this paper uses structural equation model and GSCV-XGBoost algorithm to empirically analyze the relationship among individual risk perception,leisure consumption emotion and leisure consumption behavior.The main research work and conclusions are as follows :(1)Construct the maximum likelihood structural equation model to explore the relationship between risk perception and emotion.Seven risk perception-emotion indicators were selected,and the path was designed based on TPB theory.The maximum likelihood method was used to calculate the path coefficient,and the internal quality and overall fitness of the model were tested based on 11 indicators such as factor load and SMC.The results show that the stronger the degree of stimulation of leisure consumption emotions,the higher the degree of individual attention to events,and the stronger the degree of risk perception.When the emotional dimension represents different evaluation cores,the difference leads to different evaluation tendencies of individuals,which in turn positively affects risk perception.(2)Classifier selection of leisure consumption prediction model.On the basis of XGBoost algorithm,combined with grid search and 5-fold cross-validation,the model parameters are optimized and trained,and the GSCV-XGBoost decision behavior classifier is established.Based on traditional algorithms such as decision tree and random forest,the classifier is established and compared.The results show that the GSCV-XGBoost classifier has the best effect.Therefore,this paper chooses GSCV-XGBoost algorithm to construct leisure consumption prediction model.(3)Based on the prediction model of leisure consumption level,explore the structural characteristics of leisure consumption expenditure.Based on the GSCV-XGBoost algorithm,the risk perception intensity and leisure consumption emotional characteristics are sorted according to the information gain rate and coverage rate.It is concluded that leisure consumption willingness is the most powerful factor reflecting leisure consumption,followed by risk perception.The impact of leisure consumption ranks in the top five.Age,leisure consumption attitude,perceived positive utility also have an impact on most types of leisure consumption.Perceived behavior control has a greater impact on tourism communication consumption.(4)Based on the leisure consumption location prediction model,explore the location preference characteristics of leisure consumption expenditure.Based on the GSCV-XGBoost algorithm,the risk perception intensity and leisure consumption emotional characteristics are sorted according to the information gain rate and coverage rate.It is concluded that income has the greatest impact on location selection,followed by leisure consumption willingness and risk perception.In addition,older people pay more attention to health care consumption.Compared with the existing literature,the main innovation of this paper is to explore the relationship between individual risk perception,leisure consumption emotion and leisure consumption behavior,and to combine cross validation,grid search and other methods with XGBoost algorithm to establish GSCV-XGBoost decision behavior classifier,which provides a new perspective for the study of leisure consumption.The research conclusions of this paper are applicable to the leisure consumption market when emergencies persist,and provide reference for relevant government departments to formulate policies and the rapid recovery of the leisure consumption industry.