2017-10: Implement Spam Filter for Suspicious Incoming Applications

Request No.2017-10
Date of RequestOctober 10, 2016
RequesterCCC Tech Center
Application(s)Standard, BOG, IA
Section / Page


Steering ApprovalAPPROVED
Web Service & ML data model - June 2018
User Interface to Administrator - June 2018
Steering Hearing DateNovember 4, 2016
Proposed Change to Download FileNo
Proposed Change to Residency LogicNo

See Additional Change Requests submitted while research is in-progress

Problem / Issue

Colleges have reported a rash of suspicious incoming applications from countries such as Russia and China since June 2016. These applications appear to be fakes but it's hard to discern due to some of the reasons listed below:

The problem of course, is that this week they are from mail.ru and next week they are from another email service in another country.  It's very easy to mask your identity on the internet and they will change tactics to get around any simple filtering methods.
The only way to attack the problem is to continuously learn what these applications look like and attempt to block them, just like with junk email.
As you can imagine, this is not a trivial feature.  We understand the growing concern.  I've asked our data scientist to work on it.

Proposed Solution

Investigate what it would take to build an Application spam filter where A/R staff could mark an application as suspicious and via the magic of machine learning we could build an ever updated model that could flag suspicious applications and put them in a suspense file for review.  Very similar concept to email spam filtering.


Several colleges have reported this issue and in July a special Steering Committee meeting was held to discuss the implications of these applications. It was determined that until we know more about the origin and nature of these applications it is very hard to discern if these are suspicious, fakes, spam or legit applications. Steering determined to gather more information from colleges and revisit a proposal for investigating what can be done about this - if anything - during the Fall 2016. 

Additional notes will be added here, as well as a transcript of the Steering meeting from July 2016.

Supporting Documentation

See incoming email from John Wright, West Hills College, <JohnWright@whccd.edu>:

"Reaching out to see if your team has made any progress in addressing the intentional fraudulent applications we are receiving from CCCApply. 

At this time we are close to 1,000 “bad” applications that we have caught and disabled all access to the systems. 

Because of the processing with CCCApply, we have limited tools to trace connections and find additional “network” data about the problem applications.  What logs and resources can we leverage from your team to help us trace the data? 

Any help you can provide would be wonderful.