Date of Completion

Fall 12-15-2023

Thesis Advisor(s)

Dr. Joseph Johnson

Honors Major

Computer Science

Disciplines

Artificial Intelligence and Robotics | Computer Sciences | Data Science | Mathematics | Theory and Algorithms

Abstract

This paper presents a comprehensive approach to predicting future stock prices of companies using machine learning and time series analysis. The research problem is centered around addressing the complexity and emotion-driven nature of stock investment decisions. To create an objective determinant in stock decisions, we propose a machine learning model utilizing time series data from major companies, including Amazon, Apple, Google, Nvidia, Meta, Tesla, Salesforce, Intel, and Microsoft. We explore the use of Long Short-Term Memory (LSTM) neural networks, to capture the temporal dynamics of stock prices. These models are designed to process sequential data, maintaining short term and long term information, which is crucial for accurate forecasting in financial time series. In addition to deep learning methods, we examine Gradient Boosted Trees (GBT), with a specific focus on XGBoost. XGBoost's efficiency and scalability make it an ideal choice for modeling stock price movements. It uses a series of weak learning trees to progressively correct errors from previous models, making it highly effective for regression tasks like stock price prediction. Furthermore, we delve into the Seasonal Auto-Regressive Integrated Moving Average with Exogenous Regressors (SARIMAX) model. SARIMAX extends the ARIMA model by incorporating seasonality and exogenous variables, making it well-suited for time series data with seasonal patterns. We implemented two models each for LSTM and XGBoost, one for predicting a single stock and one for predicting multiple stocks. We implemented one SARIMAX model for predicting a single stock. The models are trained and tested on historical stock data, aiming to forecast closing prices accurately. This research contributes to the study in applying machine learning to financial use cases by providing robust machine learning solutions to the complex problem of stock price prediction.

Download

Share

COinS