Depression detection using hybrid deep learning algorithms

This work is licensed under a Creative Commons Attribution 4.0 International License.

This is the repository for automatic detecttion of depression using deep learning algorithm

1. ECSMP dataset can be found and downloaded here: https://data.mendeley.com/datasets/vn5nknh3mn/2
2. "CODES_ALL" folder contains all the codes written in Python using Jupyter notebook
3. "Final_models" folder contains the final algorithm codes: hybrid models CNN-BiLSTM and VGG16-BiLSTM
4. Exploratory_data_analysis_final.ipynb is the EDA code part of the project
5. Extracting_and_labeling_filtering_signals_ECG.ipynb is the code for extracting raw electrocardiogram data from the source
6. Extracting_and_labeling_filtering_signals_EDA.ipynb is the code for extracting raw electrodermal activity from the source
7. SDS_clean.xlsx has the corresponding depression scores using Zung's Self-Rating Depression Score (SDS).