This is the experiment for Federated Online Clustering of Bandits (UAI, 2022).
├── data_preprocessing // preprocessing of data in movielens and yelp files
├── environment // environment feedback for experiment and etc.
├── experiment // main part of the experiment
├── plot // the plot functions
├── result_processing // processing the experiment result to get the desired information for plot
└── raw_data // the data files input in the experiment, which have been preprocessed
├── movielens_data
├── synthetic_data
└── yelp_data
- In CDP-FCLUB-DC experiment, we choose beta_scaling = 0.005,
$\alpha$ = 1.5,$\delta$ = 0.1,$\epsilon$ = 1. Under this condition, the experiment has achieved good results
Haoru Zhao - [email protected]
If you find any bugs or have any question, please contact me.