We fetch historical data of multiple indices and financal products. Apply alorithms to determine their market efficiency and predictability.
With various trading products, we select those with high predictability and apply machine learning and data mining techniques to build forecasting model.
Base on theories and well-trained models, we design serval trading portfolios to analyze the performance of our models.
We have done extensive research on the possibility and correspondingg methods on predicting the market. Economics theories and machine learning algorithms are studied through this phase.
We have collected data from open sources like Yahoo! Finace to prepare for the data analyses. Also, we have also prepared scripts to fetch real-time data avaible on the internet.
Based on the historical data we have collected, we use several simple methods to test the performance of different proposed methods.
We have done the first presentation regardingg the backround, proposed methods, future work of the project.
This project is built under the reqiuremnet of HKU course COMP4801 Final Year Project.