Research On Financial Conversion Rate Of New Internet Financial Users

Since the beginning of the 21st century,mobile and network technologies have entered a stage of rapid development.All aspects of our lives have been affected by Intermet technologies.The emerging cross-cutting Internet finance has also developed rapidly in recent years due to its good prospects.Various types of financial management apps have emerged like a well,and market competition has become increasingly fierce.At the same time,as financial apps are gradually saturated,acquiring new users is not only costly but also more and more difficult.After new users download applications,a high percentage of people do not manage their finances.Great loss,so paying attention to the new user financial transformation has great practical significance.New users entering the financial platform will experience several proeesses such as registration,real name,binding bank card,and first financial management.The above links are funnel-shaped,that is,each process will have a large part of users lost.The formula angle can be expressed as:new user financial conversion rate=registration to real name conversion rate × real name to binding bank card conversion rate × binding bank card To financial conversion rate.This paper is mainly planning to establish whether to predict the financial model and its influencing factors after the novice users bind the bank card,that is,the last link of the new user conversion.The model is built to include the following components:First,clarify business objectives and clean up missing and noisy data.Secondly,exploratory description and analysis of the cleaned data,including:registration time point,whether work day registration,registration month,real name time,time of card binding,analysis of the impact of data on financial management,in order to verify the stability of the model,The data is randomly divided into training set and test set;again,combined with the actual meaning and other characteristics of the model,the stability generalization of the subsequent method is improved;then,using logistic regression,random forest and xgboost algorithm,Each wealth management user models and predicts financial management after the card is tied,and sorts and analyzes the factors affecting the financial transformation of new users.Finally,the model accuracy,AUC,KS,stability and model running time are calculated,and the results are evaluated.The xgboost model was found to be the most powerful,stable,and fastest.The implementation of the xgboost model is the best and fully meets the enterprise big data scenario.It is a good new user conversion rate exploration solution,which can help enterprises save operating costs,improve revenue,and also open up the application field of xgboost.