Finding a person - A look at building a person searching algorithm
I did a presentation for the Gwinnett Microsoft Users Group on finding a person and building a custom algorithm to match incoming data to existing data.
Attached is my slide presentation. Finding a Person
The sample code I presented is on Github: https://github.com/RahnL/FindPersonDemo
Thanks for allowing me to speak. I had a great time and I appreciate the questions and feedback I got.
For me to followup on:
Are datasets slowing me down? Should I store data some other way after I get it out of the database that provides better performance?
Is it worth broadening my returned data, and paring it down in memory instead of doing repeated queries?
Would it work to hash the data, and search through the hashes? Would that be any faster? Given what that would take, might it be better to move to a NoSQL db, as I’ve been contemplating?
A note of clarification. On the records I receive (200-250K a day), a good number of them are throw out before they get to the point of hitting the person search process.
Thanks for the other feedback and questions.