Research On Financial Sequence Prediction Based On Hybrid Neural Network

With the development of China's securities market and the maturity of technology,quantitative investment has become one of the important tools for fund managers to make decisions.In China's stock market,quantitative investment strategies based on computers and big data are gradually emerging.Quantitative models have increasingly become the mainstream tools for understanding market investment.At the same time,the stock market itself has the characteristics of non-linearity and complexity,which makes it difficult for traditional investment strategies to meet people's expectations.Artificial neural network has strong non-linear approximation ability and self-learning,self-adaptive characteristics,which is increasingly prominent in the stock market prediction.This paper summarizes the development of quantitative investment in China and abroad,and the development process and status quo of quantitative investment model.Finally,the future development prospects of quantitative investment in China are analyzed and prospected.This paper analyses the drawbacks of ordinary transaction model and the greater data noise on high-frequency transaction.Even if the ordinary reinforcement learning model still can not converge well,we improve the dominant function of reinforcement learning,construct a two-network machine learning model,and combine the deep learning and reinforcement learning to make the model have a good effect on the prediction of financial data at the minute level.