Prediction Model And Investment Strategy Design Of High Send-to-Transfer Of Listed Companies Based On Integrated Learning

In recent years,the rapid development of China's securities market has led to the emergence of a number of subject stocks.According to different classifications of major events,they can be divided into asset restructuring plate,WTO plate and new energy plate,etc.Among these many subject stocks,Gao Seng Zhuan is undoubtedly the object of strong pursuit by small and medium-sized investors,but at the same time,it also produces some market chaos,which leads to investors' blind speculation about Gao Seng Zhuan stocks.At this time,it is of great practical significance to accurately predict the listed companies that may implement the "high transfer" behavior in the next year,and to screen out the corresponding target portfolio to design an efficient and stable investment strategy.This paper attempts to make some in-depth discussions on the prediction model and investment strategy design of listed companies based on ensemble learning.In terms of structure arrangement,the first chapter mainly elaborates the background,significance,contents,methods and main contributions of this study;the second chapter reviews the relevant research results of domestic and foreign scholars on the main influencing factors and investment strategies of the "high transfer" prediction model,and introduces relevant statistical learning methods,including feature selection,logistic regression model and support vector.Machine SVM algorithm;Chapter 3,starting from the analysis of the status quo of "high delivery and transfer",mainly analyzes the definition of high delivery and transfer,the current dividend distribution and transfer process of Listed Companies in China and the review of the basic situation of "high delivery and transfer" over the years,to pave the way for further empirical analysis and provide a realistic case basis;Chapter 4,starting from the initially selected four impact factors,through data preprocessing,including Data missing value and standardization processing,then using the feature selection based on tree and recursive feature elimination to screen out the six main factors affecting "high transfer",and finally innovatively using the integrated learning model of logistic regression and support vector machine(SVM)to provide a high accuracy and low volatility scheme for predicting the next year's high transfer;Chapter 5 draws from the above integrated learning model.The results show that it can achieve a higher return on excess returns.Chapter VI draws the conclusions of this study,and points out the shortcomings of the study and future research directions.The main conclusions are as follows: the investment period of the "high transfer" event in China is relatively fixed,generally about two months before the announcement of the "high transfer" plan,and the event has a strong predictability;through the analysis of the inherent law of the "high transfer" market,the portfolio of the "high transfer" forecast target stocks will be tested out of the sample every year,and the portfolio yield effect will be obtained.It is obviously better than the yield of Shanghai and Shenzhen 300 index in the same period.