How to approach this type of problem

#1

as many of the data is missing and plz tell me how to approach this problem?

#2

hey @Jan19LPN0013 ,
Generally in such case there are some conditions based on which we make decisions :

  1. if less than 20% data of a column is missing , then we impute that data using various other techniques like imputing mean values , median ,or mode values , using different imputers.
  2. if more than 50% of data is missing , then imputing it will be useless. So we just delete that column.
  3. if its between , 20 - 50 %, then we check it correlation with target variable , if it is high then we impute it else drop it. Its basically our choice.
  4. Or you can directly fill any value in that missing record and use it for training.

I hope this would have helped you to understand the major task for missing values.
Thank You and Happy Coding :slightly_smiling_face:.

#3

ohk thank you sir i will start using these tips.