This project was made so that I could make 1 billion quid 🤑🤑🤑. It involves a recurrent neural network that takes in the past 100 candelsticks of information to predict the next close price.
Make sure to checkout the notebook to see how it was made and how it works!!
- Objective: Predict the next close price of a stock using the past 100 candlesticks (OHLCV data).
- Extension: Beyond single predictions, the model forecasts several future prices iteratively.
- Explored time-series analysis and the intricacies of LSTM networks.
- Practiced feature engineering and iterative forecasting techniques.
- Emphasized the importance of data preprocessing for model performance.
- TensorFlow & Keras: For the neural network model.
- Pandas & NumPy: For data manipulation.
- Matplotlib: For plotting predictions vs. actual prices.
- Scikit-learn: For data scaling.
I think this project stands to prove a potential for neural networks in financial predictions. I also think that my model could definitely be improved upon as it feels simple and the predictions it makes feel almost trivial, e.g. going in direction in a straight line. I believe it is possible to train an AI to understand the deep underlying trends better than I have and will definitely be a project for the future!