Closed clbarnes closed 3 years ago
Multiply the distance-derived weights by some already-known factors.
i.e. instead of doing 1/(distances+epsilon), do factors/(distances+epsilon) and then normalise.
1/(distances+epsilon)
factors/(distances+epsilon)
Released in v0.4.0
Multiply the distance-derived weights by some already-known factors.
i.e. instead of doing
1/(distances+epsilon)
, dofactors/(distances+epsilon)
and then normalise.