For a detailed explanation of the methods used here for the cost-sensetive health dataset, please refer to: "Cost-Sensitive Diagnosis and Learning Leveraging Public Health Data"
For a detailed explanation of the opportunsic learning method, please refer to: "Opportunistic Learning: Budgeted Cost-Sensitive Learning from Data Streams"
Link to a preprocessed version of the diabetes dataset diabetes.pkl
- nhanes.py: implementation of the data preprocessing logic as well as definition of a few example datasets such as diabetes, heart disease, hypertention, etc.
- Demo_Dataset.ipynb: Jupyter notebook file to demonstrate the basic usage of each sample dataset.
- Demo_OL_DQN.ipynb: Jupyter notebook file to demonstrate a simple implementation of the Opportunistic Learning method.
- Other source files are used in the Demo_OL_DQN.ipynb.
- Download raw data files and decompress them.
- Install Python 3 and the following packages: joblib, numpy, pandas, matplotlib, scipy, sklearn, jupyter, pytorch.
- Use Demo_Dataset.ipynb and Demo_OL_DQN.ipynb to see a few examples on how to use the predefined tasks.
- Alternatively, you can expand nhanes.py to define new tasks by following the implementation logic of the provided samples.
If you find this repository useful, please cite the following papers:
- M. Kachuee, O. Goldstein, K. Kärkkäinen, S. Darabi, M. Sarrafzadeh, Opportunistic Learning: Budgeted Cost-Sensitive Learning from Data Streams, International Conference on Learning Representations (ICLR), 2019. Paper
- M. Kachuee, K. Kärkkäinen, O. Goldstein, D. Zamanzadeh, M. Sarrafzadeh, Cost-Sensitive Diagnosis and Learning Leveraging Public Health Data, 2019.