..-
help for ^distan^
..-
Calculates similarity and dissimilarity measures
- ------------------------------------------------
^distan^ varlist^, st^andar ^di^stan^(^distance-measure^)^ ^av^erage
^sa^ve^(^filename^)^
Description
- -----------
^distan^ calculates similarity and dissimilarity measures between
observations
Options
- -------
^st^andar standarizes the variables in ^varlist^.
^di^stan^(^distance-measure^)^. Euclidian distance is used as default.
The following
measures are allowed:
^co^variance: covariance between observations.
^cor^relation: correlation coefficient between observations.
^ma^nhattan: Manhattan distance between observations.
^mi^nkowski^[^n^]^: Minkowski metrics of dimension ^n^. Euclidian and
Manhattan distances are particular cases of dimension 2 and 1
respectively.
^average^ calculates mean distances if Manhattan, Euclidian or Minkowski
distances are specified. This option is specially useful if missing
values
are present in the original dataset.
^sa^ve^(^filename^)^ saves the results to ^filename^
Remarks
- -------
Saved output contains three variables: ^elem1^, ^elem2^ and ^distan^.
The first two
identify the observation numbers of the original dataset for which the
distance
^distan^ was calculated.
Examples
- --------
. ^distan x y z, st^
. ^distan x y z, di(mi[3.2]) av sa(c:\dist.dta)^
. ^distan x y z, di(cor)^
Author
- ------
Jose Maria Sanchez Saez