Research Of Portfolio Investment Strategy Based On High-frequency Trading

High-frequency trading is rising in recent years in our country, but profit this year is notoptimistic as well as changes in the domestic stock market microstructure. With participantsswarmed into the market, the High frequency trading is getting more difficult and the profit isdeclining. Aiming at this problem, the proposed strategy is to optimize revenue through Kellystrategy, and this is the Innovation of this paper.Modern portfolio theory are mostly based on the mean-variance model whichMarkowitz proposed in1952, this article is from another perspective, investing based on the‘fastest-growing proportion of investment funds’ theory which raised by American Bell Labsengineer Kelly.First, we will introduce the main portfolio construction theory and the development andinnovation of it. Then we will introduce the research status about Kelly criterion on domesticand foreign and the necessity to build a portfolio strategy based on Kelly optimal growthpath.This is the initially discussion about Kelly criterion.Next, we specify the Kelly strategy. First, starting from a single asset investment withbinomial distribution, we verify that its long-term optimality and analyze its nature and risks.The Kelly criterion portfolio under generalized equation gives expression to calculate theproportion of Kelly optimal investment strategy. And gives the return on assets subject toindependent and Kelly portfolio strategy portfolio under the same binomial distribution,solution and analyze its properties. Then leads to multiple problems from the investmentportfolio and build strategy based on multi-phase Kelly optimal portfolio constructionstrategies.For high frequency trading, this paper describes in detail the development ofhigh-frequency trading in our country current situation: the main characteristics ofhigh-frequency trading, high frequency trading systems, and high-frequency trading models:moving average model, Bollinger model and RSI models. Proposed a combination ofinnovative high-frequency trading strategies and methods are given to solve the model andsolving two separate high-frequency trading strategies.The final chaptet is empirical analysis. we select five minutes of CSI300index futures data, this different high-frequency trading strategies models in the same as chapter three.First we verify the advantage of Kelly criterion over a single high-frequency trading model,and then verify that the feasibility of portfolio strategy build up by two high-frequencytrading model.Finally, the work is summarized and the direction of future research is discussed.