The aim of this notebook is to classify correctly when a seizure appears in a 10 min EEG file.
The EEG data comes from a single electrode placed on different locations on the brain. The dataset contains recordings from different patients in different US hospitals.
- The data in each file is one-dimensional EEG data (single-channel) captured from a sensor with a +/- 175 uV analog recording range.
- The data was recorded at 512 samples per second.
- The data is raw, as in there has been no processing applied to the data beyond the analog hardware filters on the sensor (0.3-92 Hz bandpass).