Capstone project for the course "Deep Learning" as part of the master "Data Science and Society". Paper title "An LSTM Approach to Predicting Ethereum’s Price Trend Based on Blockchain Characteristics".
I obtained a course grade of 10/10 as a result of this paper. Mind you, this does not imply that the result/code is meaningful in any way, as the grade was solely based on the final paper.
- Use "hyper-parameter_optimization.ipynb" to see how hyper-parameter optimization was conducted. "testing_models.ipynb" shows the error metrics as well as performance of the trading strategy.
- The final results can be found inside this repository in the document going by the name of "Final Paper.pdf".
- Please don't actually trade on the results found in this paper. There is a (high) chance that applying the models I tested will generate losses in the real-world, as I didn't factor in transaction costs/slippage (and it is probably overfit).
- There is a chance that there are still mistakes present in the code.