How are you guys filling in the missing data? - Loan prediciton

missing_values

#1

Hi Guys,
I have seen that there are lot of missing data especially Credit History. Any idea how you guys are dealing with missing data?


#2

try reading about package MICE in R.


#3

@Shravanbm - There is various method by which you can fill the missing value.One of the methods is suggested by @divye_gupta.For more methods you can see this discussion.
http://discuss.analyticsvidhya.com/t/how-to-handle-missing-values-of-categorical-variables/310

Hope this helps!

Regards,
Hinduja


#4

For categorical, Build a bayes model, random forest model. These work better.


#5

Hi @Shravanbm, go through this video you will know the techniques that can be applied for missing data try all of them and see the variation in your result that will help you to have the grasp of it.Hint : Try imputation, as credit history is very significant variable in this particular question.


#6

Hi @Shravanbm

Apart from all the methods mentioned here, you can use MissingDataGUI to compare and explore :slight_smile: and go through this great article/tutorial 23 pages but worth it after this MICE and mi will have no secret for you :slight_smile:
Article link

Have a good time

Alain