A Deep Neural Network Based Model for Jane Street Market Prediction

The price trend of stocks directly affects the economic interests of investors, and also affects and reflects the country’s macroeconomic policies, so it has received widespread attention. The formation and fluctuation of stock prices are not only restricted by various economic and political factors, but also influenced by investment psychology and trading techniques. But in fact, stock prices are not only closely related to the internal financial status of listed companies, but also related to the overall stock market conditions and even the overall economic operation. Due to the many factors that affect stock price fluctuations, using traditional regression analysis models to predict is not only complicated but also less accurate. In this paper, we propose an algorithm based on deep neural networks to build stock prediction models. This model is better than other models in predicting stock trends. Specifically, the return value of our neural network model is 3137 higher than the Xgboost algorithm and 4692 higher than the Lightgbm algorithm.