Name	Name	Last commit message	Last commit date
parent directory ..
README.md	README.md
analysis.html	analysis.html
analysis.ipynb	analysis.ipynb
autoencoder.py	autoencoder.py
centralized.py	centralized.py
requirements.txt	requirements.txt
roc_delong.R	roc_delong.R

Name

Last commit message

Last commit date

Analysis of Results

Download and transform a subset of the Icentia11k dataset into the required format.

Requirements

pip install -r requirements.txt

Export the results from the mobile devices to this location, e.g., into a ./results folder.
Optional: To compare with a centralized approach, please run the ./centralized.py script first. The results (embeddings, residuals, and centrally trained model) will be saved in ./results/centralized/. Note that this requires the transformed data – the same data deployed to the mobile devices – to be available in the ./data/transformed_data/ directory.
Open the jupyter notebook ./results.ipynb

jupyter-lab

Run the notebook step by step to gain insights into the performance of the approach. The results are stored in a ```media`` folder inside this folder and must be created beforehand.