Trading Strategy Creation & Optimization with Deep Learning

Deep learning has been widely used in hedge funds and asset management firms. The increasing complexity and vast amounts of financial data have led to the need for advanced computational methods to improve trading decision-making processes. This paper presents an exploratory study on creating and optimizing trading strategies using deep learning techniques. The research aims to demonstrate the potential of deep learning in creating profitable trading strategies, focusing on using Long Short-Term Memory (LSTM) to predict stock movement and optimizing traditional momentum strategy with Deep Neural Network (DNN). The paper uses Python to create neural network models and backtest the deep learning-based strategies. The first part of the paper covers the introduction of preliminary analysis of the stock returns of JPMorgan Chase & Co.; the second part concentrates on fitting LSTM to the dataset and creating an LSTM-based trading strategy; the third part optimizes classical momentum strategy with DNN; the last part compares the strategies with the traditional buy-and-hold strategy purely in terms of cumulative returns.Bachelor of Scienc