Research On Quantitative Trading Stock Selection Strategies Of Thirty-three Degree Capital Management Company Based On Machine Learning

Private equity funds have been the first embracers of new technologies for 50 years,from the development of traditional investment theories to the combination of computers for quantitative transactions.With the gradual expansion of China's economy and market,the size of private equity capital management companies in China had reached CNY19.07 trillion by September 2019.However,compared with the international market share of 80%,the domestic market still has a lot of room for development.As far as the trend of industrial strategy is concerned,the domestic quantitative trading strategy is now in the era of combining quantitative trading and machine learning.Founded in 2015,Xiamen Thirty-Three Degree Capital Management Company has grown into a small private capital management company as the company continues to grow.In recent years,the stock multifactor strategy in the securities private fund industry has been homogenized seriously,and the profit depicted by the existing models is limited.Moreover this enterprise scale is small,the future possibly will face will lack the competitiveness to cause the loss the possibility.In the increasingly fierce competition in the industry.How to optimize the security investment fund strategy and ensure the smooth implementation of the strategy will become an important breakthrough for the future development of Thirty-Three Degree Capital Management Company.Under this background,this paper takes Xiamen Thirty-Three Degree Capital Management Company as an example to analyze the company's organizational structure,human resources status,business and effect,analyzes the current stock selection strategy of the company and analyzes the internal problems of the stock investment fund,including internal environmental problems,human resources problems,stock selection strategy defects,and compares with the domestic head private equity companies,it is found that the company needs to combine machine learning in the selection strategy to effectively solve the defects of the current stock selection strategy.Aiming at solving the actual needs of Thirty-Three Degree Capital Management Company and optimizing the stock selection strategy and its guarantee,this paper adopts the multi-factor and machine learning theory to train and test the historical data of various machine learning algorithms,and selects the best one through scoring and evaluation.Finally,the multi-factor strategy which is optimized by XGboost is the most suitable quantification strategy for the company.In addition,seven key points are put forward,including rolling training model,mixed strategy,positive attitude adjustment,timing and risk early warning,product algorithm rotation,risk control line and position control line.In terms of safeguard measures,this paper puts forward four aspects of guarantee,including system guarantee,talent guarantee,fund guarantee,technical support guarantee,and a series of optimization measures.It is hoped that this research will play a significant role in the future development of Xiamen Thirty-Three Degree Capital Management Company,optimizing the strategy of stock investment fund and enhancing the competitive advantage.At the same time,the results can provide effective reference for otherfund management companies,and help other institutions upgrade from the traditional multi-factor strategy to machine learning strategy.