Project Members: Derek Yao, Nora Povejsil, Marlon Fu
This project was made for DATASCI 207: Applied Machine Learning at UC Berkeley in the Spring 2024 semester, and explores the application of machine learning models to predict wildfire risk using satellite imagery. Leveraging transfer learning architectures, the study focuses on feature extraction and classification tasks. Through this investigation, we aim to assess model performance, identify biases, and discuss implications for real-world wildfire prediction efforts.
From the root directory, set up the environment by running conda env create -f environment.yml in the command line. The main source code for the project is included in d_yao.ipynb, n_povejsil.ipynb, and m_fu.ipynb.