Missing data retrieve from a periodic pattern.

Hi, I have a temperature record data set (attached) from 2005 to 2016. The temperature cycle is predictable because it's really just a seaosnal high and low pattern. I have the late spring, summer, and early fall record. I want to estimate the winter temperature and olug it in as a sinusoidal pattern. Can anyone help me with that?

Answer 1

[Les Beckham]

load('TempRecord.mat');
plot(T.Time, T.Temp)
grid on

There is definitely not enough information here to fill in those large gaps in the data. Perhaps you can get data from a different source to fill in the gaps.

[Ashfaq Ahmed]

But the missing data is predictable and we have an idea on how would it look like. Any synthetic data set would work as long as it makes sense.

[Les Beckham]

If you can create some sort of algorithm/equation/rule for predicting the minimum value from the data before and after, perhaps it is possible. Maybe if you had data for one winter to give an idea of the shape, you could scale/offset that data in x and y to fit each of the gaps.

However, depending on what you want to do with the data after filling in the gaps, I would be skeptical about how much you could trust the results.