4 ms·
To make it positive and to amplify it (as a side-effect). BTW, "error" is a misleading term - it communicates some fault, at least in the common sense. Distanc
by dschiptsov 10y ago
To make it positive and to amplify it (as a side-effect).
BTW, "error" is a misleading term - it communicates some fault, at least in the common sense. Distance would be much better term.
So, "squared distance" makes much more sense, because negative distance is nonsense.
- tonyedgecombe 10y agoWell it will only amplify values > 1.
- esrauch 10y agoThat's not correct. Even though the magnitudes of the value in isolation shrinks, the relative magnitudes are still amplified which is what matters. Consider values 1/2 and 1/4: in the original space it's double but in the squared space it becomes 1/4 and 1/16 so the difference is 4x. Also relevantly if you compare eg 0.9 and 1, the gap between them is amplified after squaring.
- tonyedgecombe 10y agoThose values aren't compared individually, they are summed to calculate the deviation, the result of that sum will be reduced if the values are < 1.
- klodolph 10y agoThat's a compelling case for why we should not use "distance", because distance cannot be negative, but the error term can. Just look at bog-standard linear regressions, say Y_i = m X_i + b + ε_i. It makes no sense to call the ε_i terms "distance".