EEG Schizophrenia Detection

This repository contains the code for a project on Schizophrenia Detection using EEG data. The goal is to analyze EEG data and develop a model for detecting schizophrenia based on the provided datasets.

Import Note: The majority of the logic, preprocessing, and analysis are implemented in the util module. Refer to the util module for detailed functions and processing steps.

Data

Code in this repository assumes that all data is locally downloaded and unzipped into a dataset folder. The dataset folder should include directories for each of the 81 subjects' data.

• EEG Schizophrenia Detection Dataset Part 1
• EEG Schizophrenia Detection Dataset Part 2

Note: No raw data or output files are included in this repository due to size constraints.

Project Result:

In the domain of traditional models, the Light Gradient Boosting Machine (LGBM) proved to be a robust classifier, surpassing others in this project with a ROC AUC of 95.96% and an accuracy of 90%.

Project Report

The detailed project report can be found here.