| The family is the basic unit of social and economic activities,and the active degree of household participation in investment is also an important manifestation of my country’s economic vitality.In recent years,Chinese households have begun to pay more and more attention to financial management and financial asset allocation.In 2021 The Allianz Global Wealth Report 2021 released by the European Allianz Group in October shows that after a yearon-year increase of 10.3% in 2019,the total financial assets of Chinese households will see a strong year-on-year growth of 13.6% in 2020,a growth rate much higher than the global average.In 2020,the total financial assets of Chinese households exceeded 2.5 trillion euros(about 18.6 trillion yuan),accounting for 47% of the total financial assets in Asia.This shows that Chinese households are constantly increasing their asset allocation in the financial sector,and household finance has become an important component of China’s financial structure.In addition,as household finance has become an important component of China’s financial structure,it is necessary to have theories to guide the government to make targeted financial policy adjustments,to guide the financial industry to gain inspiration from customer development,and to provide guidance for residents’ financial management.To a certain extent,the theory has this guiding role.Therefore,it is of great practical significance and theoretical value to study the participation of Chinese households in the investment of medium and highrisk financial assets and the analysis of the influencing factors.This thesis uses the China General Social Survey(CGSS)2017 data set,and after using the undersampling method to deal with the data imbalance problem,through the construction of Logistic regression model,decision tree model and random forest model to explore the identification of households participating in medium and high-risk financial asset investment in China,and then use the regression coefficient of the Logistic regression model with better identification effect and the variable importance index of the random forest model to explore the size and direction of the influencing factors of China’s participation in medium and highrisk financial asset investment households.In terms of the identification of households involved in investment in medium and highrisk financial assets in China,comparing the three models,it is found that the recall and precision rates of the positive examples of the Logistic regression model are 0.87 and 0.81,respectively,and the AUC index reaches 0.90.The full rate and precision rate are 0.88 and 0.79 respectively,and the AUC index also reaches 0.89,while the effect of the decision tree model is relatively poor.Reaching 0.80,the logistic regression model and the random forest model on the test set have relatively balanced results in the recall and precision of positive examples.Compared with the decision tree model,the results are much better,and are more suitable for identifying China’s participation in medium and high-risk financial asset investment.In terms of the influencing factors of households participating in the investment of medium and highrisk financial assets in China,this thesis uses the regression coefficient of the Logistic regression model with better classification effect and the variable importance index of the random forest model,according to the variable importance of the random forest model.And the order of importance to obtain the total family income,highest education level,personal annual income,household registration status,English level,whether the city is located,Internet usage level,whether to use online payment,Mandarin level,province where you are located,and social 18 indicators such as the importance of security projects,whether the family owns a family car,age,number of children,active political participation,mother’s highest education level,father’s highest education level,and housing area,etc.Whether the family is engaged in risky financial asset investment,The degree of explanation for the impact of activities exceeds75%.From a geographical point of view,it is found that the importance of variables in four economically developed provinces,Shanghai,Beijing,Guangdong Province,and Zhejiang Province,occupied the top four,indicating that families in economically developed provinces generally need more variables.Compared with the households in the economically underdeveloped provinces,the willingness to invest in medium and high-risk financial assets is higher.From the nature of the household registration,it can be found that households with rural household registration have lower willingness to invest in medium and high-risk financial assets than households with urban household registration.,indicating that households in rural areas are generally risk-averse.In addition to geographical and hukou factors,factors such as the highest level of education,personal annual income,Mandarin level,total family income and internet usage are factors that affect Chinese households’ participation in medium-and highrisk financial asset investment.relatively large impact.Therefore,if you want to develop a customer identification system in the financial field,you can choose and test the Logistic model and the random forest model first to save upfront costs;in addition,if you want to increase the investment participation rate of high-risk financial assets in Chinese households,apart from the geographical and account factors In addition,some measures can be taken from three aspects: personal factors,family factors,and social factors.First,in terms of personal factors,we can take education and popularize basic financial knowledge;reasonably increase the personal income of residents and continue to strengthen the promotion of Putonghua.Second,in terms of family factors,we can take measures such as raising the level of family income,increasing the employment ratio of the labor force within the family,reducing unemployment through multiple means,and continuing to improve the level of urbanization;third,in terms of social factors,we can mainly take measures such as Improve the popularity of the Internet,promote the circulation of financial information on the Internet,and realize measures such as lowering the threshold for participation in risky financial asset investment. |
