Performance Evaluation Of Financial Portfolio Based On Regression Of Elastic Network Quantile

The stock market is often regarded as a barometer of a country’s economy.China’s stock market started slightly later and is growing fast.Due to imperfect regulatory systems and complicated media commentary,the stock market turbulence became the norm.After the tumultuous trend,more and more.Irrational speculators joined the chasing team,causing the stock market to become more sloppy and fall into a vicious circle.This paper proposes the evaluation method of stocks and fund returns through statistical model,in order to provide investors with suggestions for selecting stocks and funds through modeling,so that funds can shake hands again and rationally,and contribute to the development of China’s financial investment.In this paper,the mean return and quantile regression are used to model the stock and fund returns.Firstly,the corresponding market style factors of stocks and funds are searched.Then,the weekly returns of stocks and funds are used as explanatory variables,and the market style factor is the explanatory variable.Establish a regression model.Because the market style factor is more comprehensive,it involves the problem of variable selection.Therefore,this paper gradually returns to the control group as the control group,and discusses the choice of market factors and the evaluation of stock and fund return ability by the quantile regression under Lasso penalty and elastic network penalty.In this paper,we use the Jensen α index and α* constructed by quantile regression to evaluate the profitability of individual stocks and funds.We discuss the income style of stocks and the investment style of funds through stepwise regression,Lasso punishment and elastic network punishment.Finally,the Sharpe index and the Sortino index are used to evaluate the weighted combination performance of stocks and funds,so as to verify whether the three models are accurate in the evaluation of the earnings of individual stocks or funds and the screening of market style factors.The empirical results show that under the three weighted investment schemes,whether it is the Sharpe index or the Sortino index,the stock and fund portfolio performance based on the elastic net quantile regression is better than the other two models,and the Lasso quantile returns second.This also indirectly proves that the performance evaluation ability of quantile regression is stronger than the mean regression,and the elastic network penalty is more accurate than the Lasso penalty quantile regression screening style factor,and the performance evaluation ability is stronger.