Synthetic data generation with CTGAN

Поділитися
Вставка
  • Опубліковано 14 гру 2024

КОМЕНТАРІ • 35

  • @MonsieurSchue
    @MonsieurSchue 8 місяців тому +3

    This is one of the most clear and easy to follow videos on generate synthetic tabular data. Thank you so much! This will help me tremendously :)

    • @next_phase
      @next_phase  8 місяців тому +1

      Glad it was helpful!

  • @nabilbettaieb5022
    @nabilbettaieb5022 2 місяці тому +1

    Nice video! Try using the PARSynthesizer in a log example. It may seem simple, but it's actually very complex to achieve good results. (And it's highly recommended in companies ;) )

  • @alphanzoskilliesummers6867
    @alphanzoskilliesummers6867 2 роки тому +2

    Thank you 😀Have been struggling to find an easy package for this

  • @sarahsalt3689
    @sarahsalt3689 Рік тому +1

    This was very helpful for my project, thank you!

  • @muhammadrasyidrosli9667
    @muhammadrasyidrosli9667 11 днів тому

    why after restart for update table evaluator... the gcolab wont connect again

  • @AbhishekKumar-jk1zc
    @AbhishekKumar-jk1zc Рік тому

    I am using it for a tabular data classification problem but it is throwing : ValueError: Failed to convert a NumPy array to a Tensor (Unsupported object type int), After model.fit, please help.

  • @FezanRafique
    @FezanRafique 2 роки тому

    looks cool, i will try it and let you know

  • @KaustavDas-o1d
    @KaustavDas-o1d 25 днів тому

    Thank you for sharing this useful video. But, unfortunately I am getting an error in the Table evaluator section. The error:
    TypeError: cdf() got an unexpected keyword argument 'local_ax'
    It would be great if you could help me address this issue.

  • @MrThespell
    @MrThespell 2 роки тому +1

    Hi @thenewphase - I tried to implement CTGAN but I'm facing this error while generation.
    ValueError: Shape of passed values is (10, 2), indices imply (10, 3)
    I tried modifying the mapping of multiple categorical data but unless I move some continuous features as categorical, the model is prompting this error while generating synthetic data.
    Do you know the reason?

    • @ravirajpawar5772
      @ravirajpawar5772 Рік тому

      I am also facing this issue

    • @ravirajpawar5772
      @ravirajpawar5772 Рік тому +4

      Check columns in which if any rows are blank....Put some value in it and then try ....It worked for me

    • @hiraabsarkhan6552
      @hiraabsarkhan6552 9 місяців тому

      @@ravirajpawar5772 THANK YOU SOOOOOO MUCH

  • @xkxine
    @xkxine Рік тому

    Hey! I just found this video because I am looking for some explanation about CTGAN. I have a very relevant question for me: Can i give CTGAN conditions? I dont see where i could give it some input conditions such that it gives me my output data. I would really appreciate an answer!

    • @next_phase
      @next_phase  11 місяців тому

      unfortunately, you cannot give conditions. It produces data exactly similar to the input data.

  • @jonatapaulino
    @jonatapaulino Рік тому

    Hey, congrats on the video. I'm trying to generate synthetic tabular data as well. With your tip I can create the field I want or are there fields already defined by the algorithm? For example, I wanted to create an emotions field and in that field store three emotions. It's possible?

    • @next_phase
      @next_phase  Рік тому +1

      Fields in your synthetic data should also exist in the original data. Otherwise, how can the algorithm make it?

    • @jonatapaulino
      @jonatapaulino Рік тому

      @@next_phase How many lines, for example, would I have to have in my original data to create the synthetic data? Would there be many? Thanks.

  • @hasrat17
    @hasrat17 Рік тому

    But for some parameters it is generatiing negative values data how to handle that. In your video also it generated negative value for charges?

    • @next_phase
      @next_phase  11 місяців тому +1

      You should remove them in a post processing step manually

    • @hasrat17
      @hasrat17 11 місяців тому

      Thank you for fast reply 😅 ,............
      just kidding you're video really helped Thanks

    • @next_phase
      @next_phase  11 місяців тому

      xD @@hasrat17

  • @qosaihammad5200
    @qosaihammad5200 9 місяців тому

    Hi dear how contact with you about problem solving?

    • @next_phase
      @next_phase  9 місяців тому

      Hello
      You can send me a message on Telegram: @moeen_v
      Or you can also book a free call on my Calendly. It is in the bio of the channel.

    • @qosaihammad5200
      @qosaihammad5200 9 місяців тому

      Thanks dear

  • @abhishektripathi68
    @abhishektripathi68 2 роки тому

    Bro how to create data using GaussianCopula, CTGAN, TVAE,CopulaGAN simultaneously

    • @next_phase
      @next_phase  2 роки тому

      You have to run them separately, but they are pretty much the same. Just check the SDV documentation.

    • @abhishektripathi68
      @abhishektripathi68 2 роки тому

      @@next_phase i can run them separately but according to my task i have to run them simultaneously by using multithreading i think but i'm not able to do 🥲

    • @next_phase
      @next_phase  2 роки тому

      @@abhishektripathi68 hmm that is actually tricky.

    • @abhishektripathi68
      @abhishektripathi68 2 роки тому

      @@next_phase ok😐 can you suggest me how to do in short

    • @next_phase
      @next_phase  2 роки тому

      @@abhishektripathi68Tbh I don't know how to do it but I will look it up this weekend.

  • @rithikkrishnan3433
    @rithikkrishnan3433 Рік тому

    Hi i need you help with something how to dm you?

    • @next_phase
      @next_phase  Рік тому

      contact me on telegram this is my id: @moeen_v