EEG from four canine subjects was analyzed. The EEG data is publicly available (and nicely packaged/organized) on Kaggle.com as part of the Seizure Detection challenge. The data was clipped into 1 second segments and classified as 'Ictal' (seizure) or 'Interictal' (between seizure). Features were extracted from the clips, analyzed for importance, and ran through a seizure detection algorithm. The performance of the algorithm was evaluated by ROC and AUC metrics.
data contains the training and testing data for only Dog 1 in the study. This is for use with the sample code. The data for Dogs 2-4 can be found on Kaggle. Download the clips.tar file and unzip. Please be aware that the clips.tar file is ~50GB uncompressed.
pics contains pictures of the figures used in the final report.
report contains the .html and .RMD files for the final report.
Final_Project.R is the raw code for processing and analyzing the data.
project-info.txt details the instructions for the final project in EPID 600.
Sample.R is the raw code for running this analysis on the sample data in data. This project must be your working directory path for the script to execute correctly, but otherwise should be ready to go out-of-the-box.