Text mining and content categorization

Matching e-mails with similar character strings

Reply
Learner
Posts: 1

Matching e-mails with similar character strings

Dear SAS Community

 

I would like to be able to link accounts with unique e-mails which are similar to each other, for example:

 

Account 1. vladimr241@gmail.com

Accoun 2. vladimr231@gmail.com

Account 3. vladim1245@gmail.com

Account 4. vladimra3333@gmail.com

 

The ultimate goal would be to create a summary table which would say that based on the example above, we are dealing withL

 

- 1 account holder (1 person responsible for creating all accounts) linked with 4 similar emails.

- Or we can summarise it as 4 accounts linked with 1 e-mail (so we are still assuming 1 person responsible for creating all accounts, but this time we are saying that four accounts were created using the same (as in almost identical) e-mail address.

 

I came across SAS pdf titled "Using Edit-Distance Functions to Identify “Similar” E-Mail Addresses" which discusses SPEDIS, COMPLEV, COMPEGED procedures. Unfortunately, things discussed there are quite vague and I prorably need slightly more basic tutorial, so I was wondering whether there is any standard query that would meet my requirements (i.e. summarise it in the above-described way) if I applied it to tens of thousands of e-mails address.

 

 

Frequent Contributor
Posts: 108

Re: Matching e-mails with similar character strings

It is a difficult question to answer , I tried Soundex function .But for ten thousands different names you have to find a different approach , I try to find and post the answer in any.

 

 

data newdata;
input Emailid$ 30.;
Emailid1=soundex(Emailid);
datalines;
vladimr241@gmail.com
vladimr231@gmail.com
vladim1245@gmail.com
vladimra100000@gmail.com
Hello@gmail.com
Val@1234567890
;
run;
Proc Print data = newdata;
run;

Ask a Question
Discussion stats
  • 1 reply
  • 319 views
  • 0 likes
  • 2 in conversation