Odin Sage was a conceptual study in using machine learning with quantitative analysis to generate portfolios of U.S. equity securities. We used genetic algorithms and other quantitative methods to construct a trading model. This model was tested on historical data and deployed in real-world trading. Our expertise was exhibited in the following areas:
Portfolio design and construction was based on several factors including a security’s capitalization and liquidity.
Trading & Execution
To execute trades, we utilized an order management system (OMS) to execute and manage multiple orders on 120+ securities simultaneously.
We developed the product concept and identified target markets of asset managers and hedge funds. For marketing, we compiled and generated monthly performance reports for BarclayHedge and prospective clients.
Data Management & ETL
AW developed automated script to source market data on 5000+ U.S. stocks and loaded the data into a MySQL database on an automated daily schedule.
Software was written to parse CSV files to extract 1.2 million rows of daily stock data and classify them into 11 S&P sectors. The resulting output was a time series of daily S&P sector weightings which was then used by the model.