Research On Machine Learning Stock Selection Strategy Based On Industry Stratification Of Gem

Quantitative investment is a kind of trading method that enterprises use electronic commerce or Internet to carry out quantitative transactions,and to issue purchase and sale orders through computer programming,so as to achieve stable income as the main purpose.It has more than 30 years of market development experience and history in overseas markets,in which the investment performance is stable,the market scale and share are also expanding,has been recognized by more and more domestic and foreign investors.Domestic Quantitative Investment started late,only more than 10 years of development,is still in a just beginning stage of development,but rapid development in recent years,since 2011,quantitative investment has become a hot market,the major institutions set up quantitative investment team,research quantitative investment strategy,a variety of quantitative fund products also began to appear.Machine learning is a multi-domain interdisciplinary,its research and application mainly involves the probability theory,statistics,approximation theory,convex analysis,Algorithm complexity theory and other multi-disciplinary.The discipline explores how computers can simulate or automate other human based machine learning activities or behaviors in order to gain access to fresh knowledge or systems skills,redesign,integration,or organization of these new knowledge or frameworks that people already have,so that they can continue to develop and improve their system performance.The multi-factor combination model is one of the most widely used and even the most mature quantitative financial investment management pricing models in the field of China;quantitative financial investment management,it is mainly based on the research of modern financial investment theories such as investment portfolio,capital asset pricing(CAPM)and Arbitrage pricing theory.In the investment process,there are differences in the development prospects of various industries,so we may be optimistic about a specific industry to invest to obtain income.Then this paper combines the machine learning Algorithm and multifactor stock selection,and studies the construction of stock selection strategy by industry in order to achieve excess return.In this paper,we select three big industries that many people favor for a long time —— consumption,medical biology,science and technology information,select gem stock to study,research traditional machine learning Algorithm SVM and new algorithms Light GBM and Cat Boost,which Algorithm performs better in the forecast,follow-up stock selection on which model to choose.Finally,Light GBM was selected in consumer industry and Information Technology Industry,and SVM was selected for stock selection in biomedical industry.