Replacing value by the mean of the two nearest neighbors

I have a column vector:

B = [20; 18; NaN; 25; 100; 15; -50; 23]

B = 8×1

20 18 NaN 25 100 15 -50 23

I need to pre-process data, so I want to:

1. replace numbers greater than a certain value (e.g. numbers greater than 40) with the average of the two adjacent numbers.
2. replace numbers smaller than a certain value (e.g. numbers smaller than -10) with the average of the two adjacent numbers.
3. replace missing value with the average of the two adjacent numbers.

Thank you!

[Torsten]

The result will depend on the order in which you perform these operations.

So order doesn't matter ?

[Ludovica Maria Campagna]

Yes, it doesn't matter!

Answer 1

[David Hill]

B = [20; 18; NaN; 25; 100; 15; -50; 23; NaN; 15; -15; 30];
idx=find(isnan(B));
B(idx)=mean([B(idx-1),B(idx+1)],2);
idx=find(B<-10);
B(idx)=mean([B(idx-1),B(idx+1)],2);
idx=find(B>40);
B(idx)=mean([B(idx-1),B(idx+1)],2)

B = 12×1

20.0000 18.0000 21.5000 25.0000 20.0000 15.0000 19.0000 23.0000 19.0000 15.0000

[Alexis]

This won't work if two consecutive values appear