Predictive Modeling in the Stock Market (Machine Learning)

Predictive Modeling in the Stock Market (Machine learning)

• Completed by a team consisting of Amy Larsen, Anthony Abushacra, Karan Dogra, Paolo Arciaga & Thotadamoole Shreenidhi.
• Utilized historical financial data from April 2014 to April 2024, focusing on 10 technology and automotive companies sourced from Yahoo Finance.

Objectives Achieved:
• Forecasted stock prices for the specified companies using historical data and predictive modeling techniques.
• Provided insights into potential future trends based on the developed models, enhancing decision-making processes.

Data Model Implementation:
• Preprocessed data, computed technical indicators, and constructed neural network models using TensorFlow's Sequential API.
• Evaluated model performance using metrics such as mean squared error and R-squared score, with provisions for K-fold cross-validation.

Predictive Power Evaluation:
• Achieved a predictive power of at least 75% classification accuracy or 0.80 R-squared across various companies, validating model effectiveness.

Data Model Optimization:
• Iteratively refined and evaluated multiple models using K-fold cross-validation, ensuring robust performance and generalization ability.
• Evaluated metrics such as R-squared, mean absolute error (MAE), and mean squared error (MSE) to optimize model performance.

Database Management:
• Established tables for each company's stock data, maintained data integrity, and enforced primary keys for efficient data retrieval.

Visualization and Benefits:
• Visualized analysis outcomes using Tableau, facilitating streamlined decision-making and enhanced risk management.
• Adapted analysis methodologies to evolving market conditions, ensuring adaptability and resilience in navigating financial landscapes.

Conclusion:
• Conducted comprehensive analysis leveraging historical data and advanced modeling techniques to empower stakeholders with actionable insights for performance enhancement initiatives.