Predicting Financial Markets with Machine Learning

Individual Project Preparation/Realisation (305AAE/306AAE)

Computer Hardware and Software Engineering Final Project

at Coventry University

📄 Documentation

• Final_Project_Dissertation.pdf - My Dissertation detailing the entirety of the project.
• Final_Project_Presentation.pdf - Presentation that presents a short overview of the project.
• Literature_Table.xlsx - A Microsoft Excel Table containing the sources used for Literature Review.
• Project_Notebook.url - A Link to a Microsoft OneNote Notebook containing the project Logbook & other notes.
• Individual_Project_Tests.xlsx - A Microsoft Excel Spreadsheet containing a log of the projects Tests with Parameters and Outcomes.

📌 Noteworthy Files

• downloadSequencedTrainingData.m - A MATLAB Function Script. Creates a Dataset directory with sub-directories for each class. Each class directory will be filled with .csv Data Points. The Comma-Separated Values files contain stock information for requested shares in a given time window. (Uses Lenskiy's function for downloading the data from Yahoo Finance).
• createLSTM.m - A MATLAB Function Script. Trains a Long Short-Term Memory Neural Network model from information stored in the Dataset directory.
• calculateEarnings.m - A MATLAB Function Script. Runs a simple simulation with the MATLAB NN model. Currently set up to trade for 3 months.
• batchProcess.m - A MATLAB Script. Runs the above 3 functions with varying parameters, enabling mass testing with a "hands-off" approach.

⭐️ Acknowledgements

• Artem Lensky for Creating MATLAB Functions that Enable Access to Yahoo Finance Information.