Match Levels

From Dedupe

Jump to: navigation, search

The Match Level is a reference to the field(s) being compared. The purpose of a matching exercise will determine the match level, and the match level is determined by the match weights. Consider the following list;

id name       company    address
-- ---------- ---------- -------------
1  Lee Jones  FD         1A High Sreet
2  Rob Smith  First Data 1A High St
3  Dave Owens Opus Inc   200 Main Rd
4  Bob Smith  RBS        The Lodge
5  Alex Locke FD         200 High Rd
6  Jane OShay Socx       9 The Avenue
  • Address Level:
We may be collating data for a mail shot and only want to send a single piece of mail to any one address (this mail item can be shared amongst the inhabitents). If this was the case we would expect our fuzzy system to match record 1 & 2 - we can then remove either record to be left with a list of unique addresses.
  • Individual/Contact/Person Level:
We may be inviting individuals to sign up for our new internet service and only want to send a single piece of mail to any one individual. If this was the case we would expect our fuzzy system to match records 2 & 4 - we can then remove either record to be left with a list of unique individuals. Refer to Talk:Match Levels not sure this one is valid?
  • Company Level:
We may be offering companies an opportunity to attend our conference and not want to invite more than one individual from any one company. If this was the case we would expect our fuzzy system to match records 1, 2 & 5 - we can then remove two of these records to be left with a list of unique companies.

Match Weights

Where multiple fields are being compared a final score will normally be given for the likelyhood of a match. This is figure is determined using Match Weights.

Personal tools
google ads