BRENT MARTIN
Data Scientist/ Business Intelligence Analyst / UI/UX Designer
Data Scientist/ Business Intelligence Analyst / UI/UX Designer
This project focuses on predicting Amazon’s stock prices using a range of machine learning models, including Linear Regression, Random Forest, Gradient Boosting, and an LSTM Neural Network. The models are trained on historical stock data and evaluated using R-squared and Mean Squared Error metrics. Gradient Boosting and Linear Regression demonstrated the best performance, delivering highly accurate predictions. The purpose of this analysis is to assess the effectiveness of these models in forecasting Amazon’s stock price trends, offering valuable insights into applying machine learning to stock market predictions.
Navigate to Explore Project Pages, View Visualizations and Dashboards, and Open Jupyter Notebooks:
PROJECT GLOSSARY:
COMING SOON – follow links to read informative blog posts:
The Linear Regression model was trained using historical Amazon stock data, utilizing key features like past stock prices to predict future stock prices. The model achieved a strong R-squared score of 0.9995, indicating an almost perfect fit for the training data. The Mean Squared Error (MSE) was around 0.4788, reflecting a very low error between the actual and predicted prices. While the model successfully captures the overall price trends, it may struggle to fully capture the short-term fluctuations. The plot compares the actual Amazon stock prices against the predicted prices to visualize this performance.
The Random Forest model was also trained using Amazon stock data, producing excellent results. With an R-squared value of 0.9965 and a Mean Squared Error (MSE) of around 3.0216, the model performed well, although it was slightly less accurate than the Linear Regression model. The Random Forest is more robust in dealing with the complexities of stock price prediction, but the slight increase in error compared to Linear Regression suggests potential difficulty in capturing some rapid price fluctuations. The model is still highly effective in predicting general price movements.
Gradient Boosting delivered the best performance in predicting Amazon stock prices, with an impressive R-squared score of 0.9987 and a relatively low Mean Squared Error (MSE) of around 1.1142. This model is particularly strong in capturing both the short-term volatility and long-term trends in the stock price data. Gradient Boosting’s ability to incrementally improve predictions over time allowed it to outperform the other models in accuracy, making it a valuable tool for Amazon stock price forecasting.
The LSTM Neural Network model, designed to capture temporal dependencies in the data, achieved an R-squared score of 0.9007. While the model provides valuable insights into the application of deep learning to stock price prediction, its performance was not as strong as the other models in terms of accuracy. This suggests that further tuning and adjustments may be necessary for improved predictions using the LSTM model.
This project focuses on predicting stock market movements using a variety of machine learning models, analyzing both the S&P 500 and FAANG stocks (Facebook, Amazon, Apple, Netflix, Google). By leveraging historical stock price data and key financial indicators, the project aims to forecast market direction and provide valuable insights for potential trade entries and exits.
The project applies advanced machine learning techniques such as Linear Regression, Random Forest, Gradient Boosting, and LSTM Neural Networks to predict stock prices. The models capture both short- and long-term trends, offering predictive insights that could be valuable to traders and investors.
A key part of the project involves the FAANG stock analysis, which leverages the same machine learning models to predict the price movements of some of the most influential companies in the market. By focusing on these stocks, the project demonstrates the application of machine learning to high-profile assets and highlights potential future market trends.
Client : N/A (Personal Project)
Date : September 2024
Category : Financial