Problem using isoutlier with type double

Hi guys, I am using R2013a version and I am trying to use isoutlier for my A = 1x1004 matrix (double). But it returns an error message of:

Undefined function 'isoutlier' for input arguments of type 'double'.

Error in Fringe (line 25)
B = isoutlier(A);

Is there something very obvious that I am not seeing???

Thanks!

Answer 1

[Star Strider]

‘Is there something very obvious that I am not seeing???’

Not necessarily obvious. The current online documentation is for R2017a. The isoutlier function was introduced in R2017a.

[Catherine]

That makes sense. Is there another function that I can use to find the outliers?

[Star Strider]

I’m not aware of any for R2013a. All the ones I would refer you to were introduced in R2015a or later.

With your (1x1004) vector, you can use the normal distribution. One method would be to remove all those outside of the 99% confidence limits for your data. Calculate the standard error of the mean as:

SEM = 1.96*std(vector)/sqrt(length(vector));

The data lying outside the mean ±1 SEM you could consider to be outliers, and exclude. That would be my initial approach.

There are more sophisticated algorithms available on the File Exchange, several of which are here (link).