Interpolate include NaN data.

Data likes this

-----

3

NaN

NaN

7

NaN

NaN

2

NaN

NaN

19

NaN

NaN

12

-----

How interpolate NaN data?

I try interp1 fuction. But I failed.

Please help me.

[Jan]

Whenever you post "failed" in the forum, add your code and explain the error.

Answer 1

[John D'Errico]

Download inpaint_nans from the File Exchange. Although interp1 would also have been able to solve the problem too. inpaint_nans is far simpler to use for this though.

[Jeong_evolution]

Could you make a function used interp1? I feel difficult inpaint_nans. sorry

[John D'Errico]

Why is one line of code difficult, with NO extra arguments needed?

x = [3
 NaN
 NaN
   7
 NaN
 NaN
   2
 NaN
 NaN
  19
 NaN
 NaN
  12];

y = inpaint_nans(x)
y =
            3
       5.4342
       7.0427
            7
       4.4804
       1.9813
            2
       7.0339
       13.796
           19
       19.359
       16.487
           12

Are you seriously trying to tell me that is difficult to use?

[John D'Errico]

This is easier?

k = find(~isnan(x));
y = interp1(x(k),y(k),1:numel(x));

It is not difficult. But you cannot claim it to be simpler to use interp1 here.

[Jeong_evolution]

I'm sorry. I didn't work hard.

Very easy.. Thanks ^^

Answer 2

[Jan]

data = [3, NaN, NaN, 7, NaN, NaN, 2, NaN, NaN, 19, NaN, NaN, 12];
miss = isnan(data);
data(miss) = interp1(data(~miss), find(~miss), find(miss))

[Jeong_evolution]

Thanks Jan Simon. Your answer is very inormative for me. But, I have a question. What is express "~miss"? I'm sorry.

[Jan]

"~" is the NOT operator. Try this:

data = [3, NaN, NaN, 7, NaN, NaN, 2, NaN, NaN, 19, NaN, NaN, 12];
miss = isnan(data);
disp(miss)
disp(~miss)

[Jeong_evolution]

Mr.Simon. This is real data.

When I try this code used real data, there was an error.

But I don't know mean. You know?

[Jeong_evolution]

Code is

------------------------------------------------------------------------------------

AA = xlsread('TMY_original_mssing_2');

data = AA(:,1:3);

miss = isnan(data);

data(miss) = interp1(data(~miss), find(~miss), find(miss));

------------------------------------------------------------------------------------