A Deep Learning Approach to Predicting Stock Prices One Day in the Future

Dátum
Folyóirat címe
Folyóirat ISSN
Kötet címe (évfolyam száma)
Kiadó
Absztrakt

This thesis focuses on the prediction of the closing price of stocks based on the closing prices of the past 60 trading days, in which two stocks will be the scope of this thesis, Apple stock, and Exxon Mobile stock; where Apple stock is volatile, and Exxon Mobile stock is relatively stable. data preprocessing methods will be applied on the two datasets which are provided by Yahoo Finance API and then creating 4 LSTM models, 2 for each stock where one is a baseline model, and the other is the optimized model (tuned hyperparameters), and then comparing the effect of the volatility of the stock on the performance of the models and evaluating each model based on performance metrics such as adjusted coefficient of determination, Mean Squared Error, Root Mean Squared Error, Mean Absolute percentage error, And measuring the accuracy of the predictions that fall within 5, 3, and 1.5 percent from the actual stock price.

Leírás
Kulcsszavak
stock price prediction, LSTM prediction, Deep Learning
Forrás
Gyűjtemények