You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
By supplying weights via the --objective-weights flag, the user can control the relative importance of each column in the data.
However, there is no mention what the default weights are. If you are only setting a single column, what scale should I use to match the default weights? 0 to 1? 0-5? 0-10? 0-100?
Having this explicit would be helpful.
Also, are negative weights possible? This can be used to reverse the meaning of certain measures. Where diff tends to group heterogeneous criteria, a negative weight for diff would group homogeneous criteria.
The text was updated successfully, but these errors were encountered:
Partial answer: Yes, negative weights are possible. In fact, I use an --objective-weights -10 with --objective-method diff to specify a homogeneous grouping of relative value 10.
The current documentation says
By supplying weights via the --objective-weights flag, the user can control the relative importance of each column in the data.
However, there is no mention what the default weights are. If you are only setting a single column, what scale should I use to match the default weights? 0 to 1? 0-5? 0-10? 0-100?
Having this explicit would be helpful.
Also, are negative weights possible? This can be used to reverse the meaning of certain measures. Where
diff
tends to group heterogeneous criteria, a negative weight for diff would group homogeneous criteria.The text was updated successfully, but these errors were encountered: