1 | initial version |
in the uniform case, the range is {0,59}
in the non-uniform case, the range is {0,256}
if our question was: is it a waste ? - yes.
if our question was: does it matter for the distance / prediction ? - no. adding artificial zeros does not change it, as long as some distance formula like L2, HAMMING, CHI_SQUARE is used, that strictly compares per element .
2 | No.2 Revision |
in the uniform case, the range is {0,59}
in the non-uniform case, the range is {0,256}
if our question was: is it using 256 instead of 59 a waste ? - yes.
if our question was: does it matter for the distance / prediction ? - no. adding artificial zeros does not change it, as long as some distance formula like L2, HAMMING, CHI_SQUARE is used, that strictly compares per element .
3 | No.3 Revision |
in the uniform case, the range is {0,59}
in the non-uniform case, the range is {0,256}
if our question was: is using 256 instead of 59 a waste ? - yes.
if our question was: does it matter for the distance / prediction ? - no. adding artificial zeros does not change it, as long as some distance formula like L2, HAMMING, CHI_SQUARE is used, that strictly compares per element it .
4 | No.4 Revision |
in the uniform case, the range is {0,59}
in the non-uniform case, the range is {0,256}
if our question was: is using 256 instead of 59 a waste ? (of diskspace/memory/cpu)? - yes.yes, ofc.
if our question was: does it matter for the distance / prediction ? - no. adding artificial zeros does not change it .