Polyfit function - NaN

17 views (last 30 days)
Gina Carts
Gina Carts on 18 Mar 2019
Answered: Guillaume on 18 Mar 2019
I have the age and blood measurement of some data. For some people the age is missing and I marked them as NaNs.
When I used the function polyfit to calcualte the slope and intercept it returns NaN values.
How can I deal with a missing data? Ideally I would like to keep the people with missing age.
age=[29, 31, 28, 34, 28, 31, 25, 27, 29, NaN]
measure=[82.2699,65.9515,71.3140,66.6908,76.5600,71.3734,94.6844,90.4026,95.950,97.2310]
>> polyfit(age,measure,1)
>> ans =
NaN NaN
  3 Comments
Gina Carts
Gina Carts on 18 Mar 2019
The NaN was a missing information. Can I put something else instead of NaN or should I remove that person from the sample?
Torsten
Torsten on 18 Mar 2019
Does this person add information to the sample ? I guess, no. So you should remove it from the data analysis.

Sign in to comment.

Answers (1)

Guillaume
Guillaume on 18 Mar 2019
polyfit cannot cope with missing data.So, if you want to use polyfit you have but no choice to remove that missing data.
If the reason for the missing data is completely random, then it's not an issue. If the reason for the missing data is not random, then you'll have to employ statistic methods that are robust to missing data and you need to do some research.

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!