Projects and labs from the UC Berkeley MIDS DATASCI 207: Applied Machine Learning course.
Throughout the course, I completed a series of hands-on labs that cover core machine learning workflows. These exercises include experimentation with different supervised and unsupervised learning techniques, model evaluation strategies, and tuning pipelines.
For the final project, I developed a machine learning pipeline to predict both long-term trends and short-term price movements in the stock market. The workflow includes feature engineering, temporal train/test splitting, and experimentation with models like linear regression, logistic regression classification, and the long short-term memory (LSTM) architecture. Performance metrics were evaluated across time slices, and confidence thresholds were incorporated to enhance decision-making under uncertainty. The trading strategy was ultimately assessed against multiple baselines, and the final outcomes reflected strong predictive and strategic performance.
Project folder and detailed write-up
A comprehensive breakdown of the methods, results, and lessons learned is available in the README.md file inside the project folder.