Have Geek Brain

Best practices for string affinity labels (Java). Distance, sound or combo?

Part of the process requires the use of string affinity algorithms.

The results of this process will be saved and released if let's say SS_Dataset.

Based on this dataset, further decisions will need to be made.

My questions:

  • Should I apply one or more string similarity algorithms to create SS_Dataset?

  • Any comparisons between algorithms that calculate "distance" and similarity "Sounds like"?

Does one family of algorithms provide more accurate results than others? Does the combination get more accurate similarity results?

  • Can you recommend the implementations you've worked with?

My implementation will include packages from the following libraries

http://www.dcs.shef.ac.uk/~sam/simmetrics.html

http://jtmt.sourceforge.net/

+2


a source to share


1 answer


It depends best on what you are trying to do. Soundex and Minimum Editing Distance (aka Levenshtein) are widely used because they are easy to understand. They are great for when you are trying to deal with typos or tab errors. I wish I could help further, "you'll have to experiment with how well they work for your specific purpose."



0


a source






More articles:


All Articles