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
The Fairness dashboard currently shows selection rate and fpr/fnr for binary classification. It would be useful if instead of these, the displayed plot depended on the user's chosen metrics (i.e: if Recall is the chosen metric, the plot below should be for Recall)
The text was updated successfully, but these errors were encountered:
A. metric_frame.by_group.plot(kind='pie', y='Count', figsize=[12,8])
B. metric_frame.by_group.plot(kind='bar', y='Recall', figsize=[12,8])
Where metric_frame is a MetricFrame that includes Recall and Count (defined by the user). I separate the plots because for Recall for example I would like a bar chart, while for Count I would like a Pie chart. If the same customizability could be added to the dashboard (instead of default selection rate, fpr/fnr) that would be very useful.
The Fairness dashboard currently shows selection rate and fpr/fnr for binary classification. It would be useful if instead of these, the displayed plot depended on the user's chosen metrics (i.e: if Recall is the chosen metric, the plot below should be for Recall)
The text was updated successfully, but these errors were encountered: