Alternative for average when we have noise data?

217 Views Asked by At

I am trying to get the average of some values, without letting the noises affect the result. For example, if the values were like this:

4, 5, 3, 4, -85, 6, 3, 79, 3, 5, 3, -123, 4

I would like to take the average of these values without considering the ones that are too far from the majority. In the above case, values -85, 79, -123 are very different than the others and I don't want them to affect the final average. So the average should be:

(4+5+3+4+6+3+3+5+3+4) / 10 = 4

Is there a way to do this? I thought about taking an estimate Majority or Gaussian Model, but I am not sure.

Note: I don't want the resulting average to be precise. Just a good estimate is enough.