Repository containing a project for machine learning in finance university course
Link to created dataset: https://we.tl/t-pgiAJ7nrhE
Suggested code format: jupyter notebooks
It's supposed to be a python script by the end, but it shouldn't take too much time converting notebooks to a script (hopefully)
Great guide for time series analysis: https://www.analyticsvidhya.com/blog/2021/10/a-comprehensive-guide-to-time-series-analysis/
About yahoo finance api for stock prices: https://www.analyticsvidhya.com/blog/2021/06/download-financial-dataset-using-yahoo-finance-in-python-a-complete-guide/
Person responsible: Kasia Górczyńska
• Gathered and summarised
Person responsible: Jacek Jankowiak
• Gather general statistics about the provided convictions time-series
• Identify any potential problems with data, filter outliers etc.
• Prepare a master index of stock symbols and retrieve corresponding price data for each symbol (sources, e.g. Yahoo Finance or Google)
Person responsible: Mikołaj Szymczak
• a) splits data into training and test sets,
• b) trains a dummy model,
• c) performs a cross-validation and
• d) generate stats in a form of a pdf report.
Person responsible: Kasia Górczyńska + Mikołaj Szymczak (+ Jacek Jankowiak)
• Implement a simple regression model to be used as a baseline for further analysis
• Train, test and cross-validate the regression model and show that the model generalizes on the out-of-sample data
Person responsible: Everyone
• Develop, train, test and cross-validate three alternative models
• Compare performance of the models against a baseline model and against each other for the various investment horizons
• Perform stability and sensitivity analysis w.r.t baseline model and investment horizons
All done!