Currently, it is impossible to compute confidence for some rules via UI. The minimum confidence is set to 0.01, if a rule has lower confidence than this threshold, it will be filtered out, which is a practical problem in some tasks. Also, leaving the minimum confidence threshold blank and filling in the topk field does not help, apparently, there is a default value for minimum confidence applied.
Currently, it is impossible to compute confidence for some rules via UI. The minimum confidence is set to 0.01, if a rule has lower confidence than this threshold, it will be filtered out, which is a practical problem in some tasks. Also, leaving the minimum confidence threshold blank and filling in the topk field does not help, apparently, there is a default value for minimum confidence applied.