How to find a super high or low value and normalize it?

Hello,

I am trying to normalize a series of values.

value=
60 65 70 71 70 2 70 1000 4

in this case the following values are unrealistic and too high:

value(1,6)
value(1,8)
value(1,9)

By an if command you can find the high and low values. But how can you replace them by taking e.g. the average of 3 values before and 3 values after?

Is there an easy solution for this problem?

[Jan]

Which kind of "if command" do you use for finding the outliers?

How do you want the filling to work, if several neighboring elements are invalid?

Does fillgaps solve the problem already?

[KSSV]

If the ourliers are identified......replace them with NaN's and use fillmissing.