Alternative Datasets and Enhanced Strategies with Deep Learning
Fidelity Innovation Competition 2017 for SKIW & CO
This project aims to reveal correlation between unconventional data sources and stock price movements. Our model will i) improve investment research processes and ii) trading decisions. To validate our hypotheses, we utilised deep neural networks (DNN) to predict stock price movements using our proposed data sources as DNNs are excellent in pattern recognition.
Assuming that our predictive models work, they can be applied to generate insights and enhance investment decisions, thus generating alpha for the company.
These are some implementations and visualizations for Section 3.1 and 3.3 in the report. Please refer to all the files for more details.