Optimization Method Of Quantitative Strategy By Pattern Recognition

Quantitative investment is a brand-new method that combines modern mathematical theory with financial data to gain certain returns.Although it's widely used in Wallstreet for more than 50 years,it is still a newly developed concept in China.But emerging does mean it's slowly developed,since more and more investors from Wallstreet choose to invest in China market.It's difficult to design new strategy nowadays,and how to optimize traditional strategies,how to manage multiple strategies becoming extremely important.Traditional methods to optimize strategy is based on risk analysis.It uses the diversification theory to neutralize the exposure of the strategy on each risk factor.Thereby reducing the uncertainty and enhance the strategy return to risk.This article starts from another perspective.Considering make prediction of strategy by pattern recognition,mark good strategy and bad strategy on certain data segments.So,it's actively adopt risk in the controversial way,which makes it a perfect way to compensate traditional optimization method.Quantitative strategy is based on the assumption that future market conditions can be predicted by historical data.Based on this,the situation of the future market or fundamental situation should remain in a short time,which constitutes the core for this paper.In order to analyze this problem,first of all,this paper define how to mark style of market.For pattern recognition,several quantitative methods of recognition were discussed.Afterwards,some typical strategies are used to as optimize object.It is determined that pattern recognition on the graph trend can clearly predict the future benefits of this strategy.The effect varies on the moving window of recognition and strategies.Secondly,based on pattern recognition method,the author established a more general deep learning algorithm and used regression models to classify them,providing ideas for adjusting parameters and models.Finally,by construing empirical analysis,this method was applied to the actual strategies,compared the strategy optimization of multiple strategies,multiple parameters.Then sensitivity analysis was conducted to show the advantages and disadvantages for this method.More research suggestion in the future for this method is proposed as last...