How can I get rid of outlier data in matlab?

My thermocouple measures temperature in a room. However sometimes it acts weirdly and suddenly it rises from 24 to numbers like 10000.

A= [23.4 24 10000 23.7 24.1 17989]

I want to replace the terms like 10000 and 17989 with their previous neighbour element. Like;

A= [23.4 24 24 23.7 24.1 24.1]

However there are 500 of them. I need a script that makes that automatically for me. For instance a script that finds these "absurd" data and replace it with the previous "non-absurd" neighbour element.

Answer 1

[KSSV]

    A= [23.4 24 10000 23.7 24.1 17989] ;
    absurdval = 100 ;
    idx = find(A>absurdval) ;  % Here any record greater thn 100 is absurd 

    A(idx) = A(idx-1) ;

In the above nay value greater then 100 will be replaced by it's previous value. YOu can fix your absurd value.