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
This metric is checking the names of the columns as well as the dtypes of the data. The dual-purpose is confusing, especially since (a) it makes the computation more difficult, and (b) there are other metrics that look for the validity of the data.
Let's simplify the metric.
Expected behavior
Rename the metric to TableStructure
Stop checking for dtypes in the this metric. We only need to check for the column names.
We can get rid of the parameter for ignoring dtypes
The 'Structure' property in the Diagnostic Report no longer has to compute or pass the parameter in anymore
Fix the computation to be
score = (R intersection S) / (R union S)
Where R is the set of real column names and S is the set of synthetic column names.
The text was updated successfully, but these errors were encountered:
Problem Description
This metric is checking the names of the columns as well as the dtypes of the data. The dual-purpose is confusing, especially since (a) it makes the computation more difficult, and (b) there are other metrics that look for the validity of the data.
Let's simplify the metric.
Expected behavior
TableStructure
'Structure'
property in the Diagnostic Report no longer has to compute or pass the parameter in anymoreWhere R is the set of real column names and S is the set of synthetic column names.
The text was updated successfully, but these errors were encountered: