Again assuming you are processing a static file, I would tend to attack the problem this way:

• Define a regex to recognize and extract a unique UID. Something like
      my ($uid) = $line =~ m{ <uid= (\w+) }xms;
  seems like a good first guess.
• Define a regex to recognize (and, if necessary, to extract) whatever an 'occurrence' may be.
• If both a UID and an occurrence occur in a line, increment a UID count as appropriate to the occurence. This may be as simple as
      my %occurrences;
      # enter line processing loop
      # ..., recognize occurrence, extract UID from line, ...
      $occurrence{$uid}++;
      # process next line, etc.
  if all occurrences have equal weight.
• When all lines in the file are processed, loop through the hash and send an appropriate message for each UID for which the count of occurrences exceeds the specified threshold.