JIITA, vol.8 no.4 p.1022-1030 2024, DOI: 10.22664/ISITA.2024.8.4.1022
Joshua Buhr, Even Nybo, Nicklaus Campanella, Donghwoon Kwon
North Central College
Abstract: This study explores the feasibility of utilizing a neural network model to predict stock prices. The neural network model employed is a 1-dimensional convolutional layer- based model called 1D Convolutional Neural Network (1D-CNN). Historical stock data for Tesla and Disney, spanning three years from January 1st, 2018, to December 31st, 2020, is collected using the Yahoo Finance Application Programming Interface (API). The collected stock data establishes three cases for evaluating model performance. Model training is based on window sizes of 15, 30, and 60 days and random seeds range of 1 to 1,000 with 2,000 epochs and a learning rate of 1e-3. The experimental results with three cases reveal
competitive performance for stock price prediction.
Keywords: stock prices; financial forecasting; deep learning; time series analysis; stock prediction
Fullpaper: