Recurrent Neural Networks and Stock Price Prediction Project (Exam 3)

Exam Includes:

• Fetched stocks from Yahoo Finance, XGBoost classifier, Time series split, CNN, MLP, GRU, GeLU (has a greater advantage for stocks than ReLU based activation functions), ReLU, sigmoid, softmax, and tanh activations, categorical_crossentropy, and additional insight into accuracy and RMSE via log bar graphs
• Results are performances of models with their trained and validation parameters to show which model is best and worst at handling these combinations of stocks (Apple, Dollar General, Blackberry, Amazon, and S&P 500 ETF)
• Deals with combination of stocks instead of each stock individually for generalized performances, less extraneous work, and identifying common patterns/shifts within these stocks including their feature column labels along with target outside of the column labels
• Each model is over 50 epochs without early stopping to visualize a greater outlook of the performances over each epoch