2012 Kaggle Competition

For the final project of Spring 2012 Stat 151B at UC Berkeley, Jeffrey Lin, Kelly Huang, and I created a model to predict the light intensity of an image being shown to a person, based on the person's brain activity (captured via fMRI). Our best model was an average of GBM, Lasso, and k-NN models; and our predictions received the second lowest test error in the class competition.