Why this total residual error shows 0.5?

I confirm the function 'findchangepts' following code.

findchangepts([0 1 2],'MinThreshold',0.5)

this result shows total residual error = 0.5. But, in my understanding, each points of residual error are 0, -0.5, +0.5, so the total residual error is 0. Because the values of residual error is each points to red line.

Could you explain me why this total residual value is 0.5?

Answer 1

[Shunichi Kusano]

Hi Kazuki,

as described in the documentation, it is "the sum of the residual (squared) error".

Hence, in this case, 0^2 + (+0.5)^2 + (-0.5)^2 = 0.5.

HTH.

[Spreadea]

I see,I just miss understanding. Thank you very much answer for me!!