|
|
Scam in Action
Click
Here to (Safely) Check Out Email Scams In Action
SPAM - The Real Deal
Current SPAM techniques fall into a trap of assuming
that everything that isn't identified as SPAM must be wanted. In
reality, there is a "grey area" that accounts for a massive
amount of text, knowledge, information and other material that is not
SPAM, but is of no current interest to the user when it is received.
It's information that is relevant to the user, but just something
that is not directly of interest to them at the time. Current SPAM
techniques, such as Bayesian filters, take a sample of text that the
user is interested (it's termed "ham") and store it (in
something called a corpus). Incoming email is compared against the
stored "ham" and everything that doesn't match is assumed to
be SPAM. SPAM text is stored in it's own corpus.
It's a system that can
work well for individuals but starts breakdown when:
|
The user
is working more than one specific field, such as the medical
legal field |
|
The user
changes interests or has multiple interests |
|
The user
doesn't understand the dynamics behind SPAM and all this
technical stuff. That's pretty much everyone. |
|
The user
resorts to "white lists" and "black lists"
to override the SPAM detection and it opens really ugly security
problems. |
Trusted Lists. Trusted?!
Existing SPAM "solutions" allow the user to
override the SPAM filtering with "white lists" and "black
lists". If you trust someone not to send you SPAM, you can put them
in a "white list". Conversely, you can put someone in a
"black list" if you want to disregard everything they send.
Although seductively simple in their approach, these lists open the
doors to Phisher's,
bogus email, security
problems and ultimately compromise entire process. Such lists are
used because they were the only solution available and there really
hasn't been a way to resolve the bogus
email problems.
Until now that is.....
The emSorter Difference
Building on emSorter's categorization
abilities, we define SPAM as a mixture of different texts that the user
is not currently interested in. It's True! In a couple of weeks time the
user might want that special credit card or need organ enlargement.
emSorter uses a combination of Bayesian
(style) filters and the Artificial Intelligence behind "smart
search" (that also powers jukeSpace)
to classify incoming emails.
The results are stored in categories that
adapt over time to suit the users interests. We call them Self
Organizing Communities of Knowledge or SOCKS. New
interests are added and entries applying to old, unused and unwanted
interests are dropped.
Trusted Lists - the emSorter way :)
Because we can identify and trap bogus
emails, our use "white lists" and "black lists"
is inherently more secure, more reliable and since we can rely on the
content being genuine - trusted. Content from these sources can
optionally be classified to form part of the overall categorization
process.
Learn from others....
emSorter, allows you to import and share category information from
others allowing you to get a head-start on categorizing your
email and beating SPAM. By sharing classification information with
others of similar interest, an individual user benefits from the
definitions from a multitude of sources. Read more here...
|
Copyright © 2000-2004 FINDbase, LLC. All rights reserved.
Please check our
Legal Notices and terms of use.
kimSPACE, "Adaptive Stores", "because searching sucks", "self organizing communities of knowledge"
jukeSPACE and FINDbase are trademarks of FINDbase LLC.
All other trademarks are acknowledged to be the property of their respective owners.
FINDbase technologies are subject to Patent protection in the USA and other countries.
creative design by: Riavon Enterprises
