Prediction of High Frequency Trading Financial Data Using Stacked LSTMs for Algorithmic Trading

摘要

This paper is focuses on the prediction of high frequency trading time series financial data using stacked LSTMs model for investment strategies in algorithmic trading. The goal is to give a deep learning (DL) approach that can be potentially beneficial to the complex investment strategies in algorithmic trading. This paper is not a complete investment strategy that generates a profit and loss curve (PNL). Rather, it shows how LSTM based deep neural networks can be applied to predict time series financial data. It can predict the asset prices or asset returns, and the generated predictions can be used to make certain decisions such as open a long position or close the short position.