search

acphile / RTSGAN

35 stars 8 forks source link

Reproducibility of baseline models #6

Open hojjatkarami opened 9 months ago

hojjatkarami commented 9 months ago

Hello,

could you please kindly share with me how did you adapt TimeGAN and DoppelGANger to irregularly sampled time series dataset?

I would appreciate it if you have any code with me. Or any additional library if you used.

Many thanks

acphile commented 9 months ago

Sorry, I did not save the code for the baselines. You can adapt them in their corresponding repos. The idea is just following how we use RTSGAN to generate incomplete time series: Treat each missing value as an additional mask to generate and use the synthetic masks to form the final irregularly sampled time series.

hojjatkarami commented 8 months ago

Hi,

Could you share the hyperparameters for replicating RTSGAN results on stock and energy datasets?

Thanks

acphile commented 7 months ago

You can check the parameter setting in the paper: For stock, it is default in the code and for Energy, it should be --hidden-dim 112 --embed-dim 448 --noise-dim 448 --layers 3