I have 500 companies individual data for the period 2009-2021. There are so many missing values in the data. So I deleted those company whose data is missing greater than 5 percent. So with the remaining companies I waana replace these missing values. So linear interpolation is good or not. Or shall I use mice in r.
It's a good option but it will be wise to check the relative operating characteristics of the entire data when used or validate your result first to know the extent of reliability of the computed missing data before making use of such result. Meanwhile, other common properties of companies with the most missing value can also help in narrowing down your estimation, rather than working with the entire dataset. Really hope to see how this goes, Komal Jindal
Thanks for the answer but how do we check relative operating characteristics
Hello,
you can also use "NeuralPower Softwer". I had used it and it's very reliable. It is for free for one month.
good luck,
Azra