Closed alvinthai closed 7 years ago
https://github.com/mlr-org/mlr-tutorial/blob/76157380287fca598a4e564c3e700386e7ff734a/devel/html/measures/index.html#L1181
according to the definition: (number of concordant comparisons - number of discordant comparisons) / (number of concordant comparisons + number of discordant comparisons)
the limit for best score should be 1 (number of discordant comparisons = 0)
Following the logic of this formula I guess you are right. Maybe @larskotthoff knows something more. Probably a copy & paste error of the G1 index?
Yep, that's a typo. I'll fix this.
https://github.com/mlr-org/mlr-tutorial/blob/76157380287fca598a4e564c3e700386e7ff734a/devel/html/measures/index.html#L1181
according to the definition: (number of concordant comparisons - number of discordant comparisons) / (number of concordant comparisons + number of discordant comparisons)
the limit for best score should be 1 (number of discordant comparisons = 0)