Learning to smell for wellnesss

This repository contains the code to reproduce the results in the paper "Learning to smell for wellnesss". The code here replicates the result in 1(1−5) − 2 in table 1 of the paper. Same logic goes for the rest of the results.

Datasets

Datasets can be downloaded from the links described in the paper.

Dependencies

Tensorflow==1.5.0.
Keras==2.1.2.
Python==3.5.4.
scikit-learn==0.19.1.

Training and Testing

python eval.py will train the model and test the target data over 5 iterations. Note that the result may be a bit higher or lower than the one reported in the paper due to some randomness issues but should converge to that value over longer iterations.

Citation

@article{owoeye2019learning,
title={Learning to smell for wellness},
author={Owoeye, Kehinde},
journal={NeurIPS Joint Workshop on AI for Social Good},
year={2019}. }