A couple of years ago I got the farm girl spam for a little while - made me wonder where I I had wandered to get on the list.
Like others, I now get some foreign spam. From France or Japan most recently.
How are spam filters made?
I get spam and have had problems with my e-mail blocking way to much to where friends e-mails don’t come through. How are spam filters put together?
This is what i see when I open my spam filter option;
“Message filtering allows you to define rules and actions that are triggered when messages are delivered to this mailbox.”
“The following filters are enabled for this mailbox”
Then there is a box to enter in a filter.
An example of a span e-mail I just recieved along with a bunch more this AM
“From” Tina Duvall “Subject” Hair loss. Now I am not loosing my hair so I don’t need this. How would I filter it out?
[I do have administration filters enabled, “High Risk”= delete “Moderate Risk”=delete
“Low Risk”=Allow]
Thanks for any help!
There are a bunch of approaches to spam filtering.
At the mail server end, there are Realtime Block Lists and IP Block filtering. These prevent certain IP addresses from delivering mail. The block lists are maintained by people that mail administrators trust (directly or by reputation). IP Block filters reject mail that has a source address in a residential IP address block, as most spam is generated by compromised residential PCs (to avoid the RBLs). Greylisting is also used - this is where the receiving mail server asks the sender to delay delivery for a couple of minutes. A RFC-compliant sender will do so, whereas botnet mail delivery agents usually don’t - they just give up.
Large ISPs can detect spam by content analysis - if thousands of their customers receive the same (or substantially similar) messages, chances are it is spam. This is why much spam has randomly generated/selected text included.
Finally, there is Bayesian filtering, both in the mail client and at the ISP. This uses a statistical filter to score the email. This filter is built up by comparing emails you want to keep (ham) to emails you don’t want to keep (spam). With a sufficiently large corpus, Bayesian filtering can be very good. But it relies on accurately managing the ham/spam data, by clicking the “Is this spam” checkbox/button on emails.
The final local client filter is based on your address-book. If you maintain this properly, your mail client will score email from your friends as more likely to be ham and not spam.
Si
Si,
You obviously know what you are talking about, but that explanation went over my head like it was an F-16 Falcon.:o (but I still have my hair.
What kind of entry or string of words would I put into a spam filter for an example like I posted?
I know this isn’t a terrible spam, although I do not open anything I do not recognize.
Note;
I do not use any browser type of pop3 e-mail programs. The kind I use lists the information in my example and downloads only if I click on it. The problem is I have to keep deleting this garbage or my inbox on the server will fill thereby blocking e-mails I need to see.
Thanks,
I don’t know what I’m doing right but I get maybe 2 or 3 spams a month in my “real” email account and in my secondary hotmail account. If that.
Fair enough - it was a pretty high level flyby.
What mail client do you use, because they are slightly all different.
From your description, I think you have a POP3 mailbox at your isp, and a client that is set up to initially retrieve headers only. Then you can mark emails for deletion, or download them for reading. Is this correct?
If it is, it is harder to set up your filtering, because initially, you only have headers to work with, rather than the full body. But we can try, and I am basing this on the Thunderbird email client…
You need to name your filter - say NoHairLoss. You want to set the field to Subject, have it say contains “hair loss”. Then set the Action to Delete.
Using this method, you will have to spend a lot of time sorting out what filters to apply, I’m afraid. However, with something like Thunderbird, you can flag junk emails (there is a button on each message you can click) and delete all the flagged junk mail. As you do this, it get better at figuring out what messages are spam (via a Bayesian filter). This may be worth a try.
Si
Si,
Thanks for the help!
The e-mail I use is through my web site host and they now use, “© MailEnable Pty. Ltd., 2001-2008. All rights reserved.”
Whatever that is.
But I did get things figured out. i was wading to shallow when I was looking into my spam filters. I get everything I think I want when I click on my “Add Filters” button. I thought I had to have something entered into that box that looked like a text box:smack:,
Well I am going to work with this on all the e-mail accounts I have through my web host that are in use by my Grandchildren.
Thank You for helping me step a little deeper into this confusing murky water 
I am now going to strap on a gun and head to the bush and look for Antler sheds. Somewhere I am more at home.
Greg
This is not a counterexample to what I said, since even if it does block most or all spam, it’ll also block many legitimate messages. There are a lot of people who legitimately should be able to e-mail you, and if you disseminate your e-mail address widely enough to get all of them, you’ll almost certainly end up getting it to folks who shouldn’t have it, too.
Just one example: I teach labs, and my students need to be able to contact me, so I put my e-mail address on the syllabus. One of my students e-mails me to tell me that he won’t be able to make it to next week’s lab, since he has a doctor’s appointment. My e-mail address is now on that student’s computer. Now, that student gets a virus, and the virus scours his entire computer looking for e-mail addresses, and finds mine. The virus then either sends spam itself to all of the addresses it found, or reports them back to some central spam clearinghouse that does the same. The only way I could have stopped my e-mail address from getting out there in this situation would have been to not tell it to my students, but then, I’d have missed that e-mail about the student missing lab.
So use something throwaway, like carnivorousplant@isp.com to use with your students or someone likely to misuse, however inadvertently, your address. 
I don’t want to get stuff from, say HP. I lie when an email address is required for example to download a printer driver. john.smith@nowhere.com, although that one was once refused, “address already on file.”