Top Dog and Spam Hound use state of the art spam detection methods.
The history of spam detection
is full of magic bullet solutions that are effective for a short time
and then stop working.
The reason for this is that spamming methods continually evolve.
A method that works today is no guarantee that it will work tomorrow.
Top
Dog and Spam Hound
take a different approach to spam detection.
All filtering methods have their strengths and weaknesses.
Instead of relying on a single test, Top
Dog and Spam Hound
use a full array of different tests combined with a Neural
Network which learns based on your personal email usage and changing
spamming methods.
This means that they get better and better at detecting spam the longer
you use them.
Some of the methods Top Dog and Spam Hound use to detect spam are:
- Bayesian filtering.
- Cooperative filtering.
- Heuristic pattern detection.
- Blackholes (DNS databases).
- Sender address verification.
- Whitelists.
- Content filtering.
Neural Networks are modeled
on the brain.
Like a brain they have the ability to learn patterns and adapt to changing
circumstances.
This is an important ability in the face of changing spamming practices.
Bayesian filtering is one method
of spam filtering that Top
Dog and Spam Hound
use.
It performs a statistical analysis of the words that make up the email
to determine if the words are more commonly found in spam or in personal
email.
This method has become very popular lately which is why now spammers often
add a lot of random text (e.g. "Tables bark deliciously...")
to their spam.
The idea is to swamp the spammy content (signal) with random text (noise).
This is also why some spam email contain only an image with the text printed
on it.
Spam filters which rely solely on Bayesian filtering can miss these spam.
Bayesian filtering does however remain effective in detecting personal
email since spammers are unlikely to know your personal interests and
so email that contain words relating to your interests are easily detected.
Cooperative Filtering.
Cooperative filtering is where a large number of people get together and
flag email as spam.
They then share this information with one another thus avoiding the necessity
for other people to read the spam email.
To preserve privacy
Email is shared in the form of "Checksums" which are sort of
fingerprint of the email.
Top Dog and Spam
Hound use two such cooperative filtering systems: the Distributed
Checksum Clearinghouse (DCC) as well as Apocgraphy's proprietary database.
Currently more than 150 million email pass through the DCC on week days.
Apocgraphy's database is smaller but it has the advantage of storing email
hoaxes as well
as numerous other
types of email.
Cooperative filtering is extremely effective though it suffers from the
problem that people sometimes disagree on what is spam and what is not.
Spam filters that rely solely on cooperative filtering run the risk of
blocking email that you do not consider spam.
Top Dog and Spam
Hound on the other hand use cooperative filtering as just one of a
number of filtering methods and automatically learns your email preferences.
With heuristic pattern detection
spam filtering is done by looking for certain spam signs in the email
that usually indicate spam. These indicators are all weighted as to how
strongly they indicate spam and then added up to yield a "Spam score".
An example of this is a "Sent" date that is in the future. This
is sometimes used by spammers to place the spam at the top of your inbox
when you sort by date.
Historically this has been a very effective approach. However spammers
have learnt to avoid many of the spam indicators checked for.
As well, what may be a spam sign for someone else may not be a spam sign
for you.
Never the less spammers are rarely able to deliver email the way legitimate
people do and frequently want to place pictures and such in the email.
This makes it difficult to avoid spam signs altogether.
Filters that rely solely on heuristic pattern detection are generally
on the way out because they do not perform as well as other methods of
spam detection.
Top Dog and Spam
Hound on the other hand use heuristic pattern detection as a part
of an array of filtering methods and feeds the results into a Neural
Network to adapt to your personal email usage.
A Blackhole is an online database
of computer addresses (IPs).
Computers are added to blackholes for any number of reasons ranging from
belonging to a company that continuously and persistently sends out millions
of spam, to computers which reside in a particular country.
Blackholes provide very good spam filtering as part of a series of tests
especially when configured for an individual user.
For example, a result that indicates that the email was sent from a server
in China (where security is notoriously lax) is not an indication of spam
if you regularly correspond with people in China.
However if you don't, it may be a good indicator that the email is spam.
Top Dog and Spam
Hound use a Neural Network to automatically
learn what blackholes are relevant to you.
When you receive an email from
an address that you have never received an email from before Top
Dog and Spam Hound
automatically check that address to see if it is valid.
Some spam filters offer to "bounce" spam for you.
This is a bad idea for a number of reasons:
- It assumes that the return
address of the email is a valid address.
This is almost always not the case. - It assumes that once the
spammer receives a bounce they will take you off their list.
This is also unlikely since spammers know about this trick and in any case do not find it worth their while to keep their lists updated. - Often the return address is does not belong to the person who sent the email. Often a spammer will choose the email address of an innocent third party to use as the return address. As a result that person gets buried in a mountain of bounced email. By faking a bounce you are only adding to the problem
Because most people communicate
with a number of different people on a regular basis, an important part
of a spam filter is allowing you to indicate that certain senders are
pre-approved and others you never want to hear from.
When installed Top Dog
and Spam Hound automatically
scan your address books as well as your "sent items" folder
and extract the addresses of people you have corresponded with in the
past. These people are whitelisted.
In addition to this, every time you send an email Top
Dog and Spam Hound
will automatically white list both the person you sent the email to as
well as the subject line and other information about the email.
As a result you are assured that Top
Dog or Spam Hound
will not block any response to your email no matter how spammy the response
may look.
This means that the longer you use Top
Dog or Spam Hound,
the better they get at detecting spam.
In addition to whitelists Top
Dog and Spam Hound
also keep a list of people who's email you want to block.
If any spam makes it through, with a single click you can categorize the
email as spam.
Categorizing an email as spam adds the sender to your personal blacklist
and trains the Neural Network to detect similar
spam in the future even if sent using a different address.
Top Dog and Spam Hound come pre-loaded with four different lists of words and phrases that you can choose to block.
- Pornographic Phrases - This list contains words and phrases that are both sexual and crude.
- Sexual Phrases - This list contains words and phrases that are sexual but not necessarily crude.
- Profanity - This list contains swear words and phrases.
- Miscellaneous - This list contains words and phrases that are commonly found in spam.
All lists are completely optional
and you can add or subtract from them to suit your preferences.
Whitelisted senders are of course exempt from being blocked due to using
words or phrases in these lists.