Patient Stroke Prediction (Class Imbalance) - Data Every Day

Поділитися
Вставка
  • Опубліковано 19 лис 2024

КОМЕНТАРІ • 8

  • @Raj-sz9pg
    @Raj-sz9pg 3 роки тому

    Thank you Sir.
    It helped alot in terms of understanding nice explaination.
    💯

  • @peterjohngerero150
    @peterjohngerero150 2 роки тому

    Nice

  • @Mohamm-ed
    @Mohamm-ed 3 роки тому

    Thabks Gabriel for the daily effort I have a question regarding the dataset. I have an EEG dataset without the last column for example in this dataset the stroke column which is 1 or 0. How can I train this kind of dataset ( let's say 32 channels EEG data represents 32 columns only ).

    • @gcdatkin
      @gcdatkin  3 роки тому +1

      Well, what is it you would like to make predictions about?

    • @Mohamm-ed
      @Mohamm-ed 3 роки тому +1

      @@gcdatkin thanks very much for replying. I want to predict that the person has pain or not.

    • @gcdatkin
      @gcdatkin  3 роки тому +1

      @@Mohamm-ed In that case, you will definitely need a column containing the labels for each EEG reading (each reading must be labeled with 0=no pain/1=pain). Otherwise, there is no way to tell the model if it is correct or not in its prediction.

    • @Mohamm-ed
      @Mohamm-ed 3 роки тому +1

      @@gcdatkin 🌹🌹🌹 Thanks again.

    • @Mohamm-ed
      @Mohamm-ed 3 роки тому

      @@gcdatkin another question sorry for disturbing you. The EEG data is a time series do I need to lable it. I saw a code about sensor array data ut has only the sensors reading and time and then they delete the time and make a model for the sensors readings without labeling the data.