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
Tricky, especially since it's unlikely the user will have an intuition for the ''correct'/'best' parameter (it wouldn't be very good if they set it to like, 100 or something!)
I'd possibly set the default value to 1e6 which will be quick to compute and will produce 'reasonably decent' estimates. And make sure we try to ensure in examples/tutorials we always explicitly set it to sensible values
I'd possibly set the default value to 1e6 which will be quick to compute and will produce 'reasonably decent' estimates. And make sure we try to ensure in examples/tutorials we always explicitly set it to sensible values
My only concern with that is that we don't really have a mechanism for flagging that estimates may be unreliable - I'd worry some users might not look at the docs and just use the default, and then get (somewhat) unreliable estimates. I guess we could emit a warning if it's left at this value that they may want to increase the number for more serious estimation?
Splink 4:
Get the error:
TypeError: unsupported operand type(s) for *: 'int' and 'NoneType'
Some options for behaviour:
max_pairs = None
means we set proportion as 1, similar to if we have moremax_pairs
than possible record pairsexplicit error message if it is left asmaking this argument non-optionalNone
, forcing user to set a non-default valueThe text was updated successfully, but these errors were encountered: