The existing uses of SimMetrics

  • ReverendSam

    Please leave a description here and a link if possible to the project or task you have applied simMetrics to. Below are a few examples I know about already but if you have more details and a web link please add it.

    Fraud Detection
    Plagurism Detection
    Ontology Merging
    DNA analysis (student projects)
    RNA analysis
    Image analysis (used for fast feature comparison)
    Evidence based machine learning (merging multiple captures)
    MS Excel plugin for cell similarity
    Database deduplication
    Data Mining
    Web Interfaces e.g. Ajax style suggestions as you type
    Data Integration
    Semantic Knowledge Integration

    and a whole lot more

    The people behind simMetrics would really love to know more about anything you use it for please leave a msg.

    • ReverendSam

      Additional usage - fingerprint analysis, comparing binary feature vectors taken from black and white fingerprints to determine fast comparisons for fingerprint search.

    • ReverendSam

      I got the following reply about the use of simMetrics in CheeseLab (i asked what it was) a simple usage but effective see for more details.

      Well cheeselab is a mud (multi user dungeon) mapping proxy specifically for MUME (multi users in middle earth) that parses text output from the mud world and creates a real-time graphical representation of what the world looks like.  It uses simmetrics to handle situations where typographical errors are corrected in room names and descriptions during movement validation, so that if the implementors for example fix a room name from "The Screendoor" to become "The Screen Door" the mapper (using simmetrics) assumes it's the correct room because it uses the percentage of similarity to determine correctness and not just a string match.

      That's probably longer than you want but I'm very bad at writing things concisely.  I would like to say that simmetrics is awesome, one of the first problems I ran in to when making cheeselab was during the time they were correcting many typos in rooms and caused me endless headaches when my database became out of sync, but still very close to what they had written.  After some quick searching I found simmetrics and in less than 30 minutes I had it integrated in the validation routines and no longer had any problems.  The classes were easy to use and well documented and completely painless to integrate :p

      And I'll check out the new version, so far though the current version has been more than adequate, I use the Levenstein class and just compare the getSimilarity() value to a configurable percent for the user, I use a 98% similarity for names and 90% similarity for longer descriptions which seems to work great.  Thanks so much for your work and making it available for others to use!

    • ReverendSam

      oddly SimMetrics is now being used in historical record scanning to try and identify war criminals....