search

ydataai / ydata-synthetic

Synthetic data generators for tabular and time-series data
https://docs.synthetic.ydata.ai
MIT License
1.38k stars 232 forks source link

[FEAT] How can i set min and max bounds for columns with numberic data? #328

Open tsiakmaki opened 5 months ago

tsiakmaki commented 5 months ago

Our columns with numeric data are integers above or equal to zero. We noticed that the output synthetic data contains numbers below zero in those columns. Is there any way to set thresholds for the synthetic numeric data?

Thank you in advance for your effort.

AryanSaeedi commented 5 months ago

Yeah, I am getting the same issue with all of the models I worked with so far. DRAGAN, WGAN, WGANGP, WGAN, and with CramerGAN after a while I get nan for generator's and discriminator's losses. I am using a network traffic dataset and it can't have negative values. I even tried running WGANGP for 3000 epochs thinking it might try to improve itself but still, I was getting negative values. My dataset is (300000, 80) where only one column is categorical and the rest are continuous. Please let me know if you were able to solve the issue.

Thank you