(PROJECT)
FER2013 Machine Learning Coding Competition Attempt
PROJECT OVERVIEW
Development and evaluation of a facial expression classification model with emphasis on preprocessing, diagnostics, and visualization.
KEY HIGHLIGHTS
Model Development
Engineered a facial expression classifier using Python and NumPy, achieving 97% training accuracy on the FER2013 dataset.
Performance Analysis
Analyzed confusion matrices and loss curves to diagnose overfitting and identify a 42% validation accuracy gap.
Visualization Tools
Developed custom visualization tools to display predictive certainty and randomized validation samples.