Matchcode Optimization

From Melissa Data Wiki
Revision as of 16:57, 20 September 2018 by Admin (talk | contribs)
Jump to navigation Jump to search
ICON WIKI Warning.png
This page is still under construction!
Melissa Data strives to give you the most complete and up-to-date information about our products as possible. To do this, we must maintain our documentation. This means the content may not be complete or correct. Use at your own risk!

← MatchUp Hub

Matchcode Optimization Navigation
Matchcode Optimization
First Component
Fuzzy Algorithms
Swap Matching
Blank Matching
Advanced Component Types
Dice's Coefficient
Double Metaphone
Fast Near
Frequency Near
Jaccard Similarity Coefficient
Longest Common Substring (LCS)
MD Keyboard
Overlap Coefficient
UTF8 Near

Evaluate Matchcode

To help you understand the effect of constructing and implementing a sub-optimal matchcode, the Evaluate Matchcode feature evaluates the five critical areas of a matchcode which determine whether one can expect the best throughput when running a process. Evaluate matchcode will indicate if your matchcode is optimized, has a warning, or is sub optimal.

MCO Evaluate Optimized.png

MCO Evaluate Warning.png

MCO Evaluate Sub.png

Five Critical Matchcode Areas

First Component Used in All Set Columns

When the first component is not checked (not evaluated) in all used combinations, processing will be very slow, as matchkeys cannot be grouped by that first component into optimal neighborhood clusters.

Fuzzy Settings

If a component uses a fuzzy algorithm other than exact, each record compared will produce speed penalty on keybuilding or on that algorithms required computation when deduping.

Swap Matching

When each key is compared, swapping one records values (for example: an inverse name format) and reattempting to identify a match will produce slower processing speeds.

Blank Matching

Allowing one of the 2 values being compared to either be blank or a single initial will require the matchup engine to make an extra comparison(s) and will produce a speed penalty.

Advanced Matchcode Component Type

Most matchcode component types simply specify how to extract source data into a records matchkey, but Advanced types will also tell the deduping engine to perform a calculation foreach record compared, even if the Distance property is set to zero.