You can subscribe to this list here.
2004 |
Jan
|
Feb
|
Mar
(27) |
Apr
(25) |
May
(8) |
Jun
(2) |
Jul
|
Aug
(1) |
Sep
|
Oct
|
Nov
(1) |
Dec
|
---|---|---|---|---|---|---|---|---|---|---|---|---|
2005 |
Jan
(1) |
Feb
|
Mar
(1) |
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
(2) |
Oct
(1) |
Nov
|
Dec
|
2006 |
Jan
|
Feb
|
Mar
|
Apr
|
May
(1) |
Jun
(1) |
Jul
|
Aug
(1) |
Sep
|
Oct
(2) |
Nov
|
Dec
|
2007 |
Jan
|
Feb
(1) |
Mar
|
Apr
|
May
(8) |
Jun
|
Jul
(3) |
Aug
(8) |
Sep
|
Oct
|
Nov
(1) |
Dec
|
2008 |
Jan
(3) |
Feb
|
Mar
|
Apr
(2) |
May
|
Jun
|
Jul
(2) |
Aug
|
Sep
(3) |
Oct
|
Nov
(3) |
Dec
|
2011 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
(1) |
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
2012 |
Jan
|
Feb
|
Mar
|
Apr
|
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
(2) |
2013 |
Jan
|
Feb
|
Mar
|
Apr
(1) |
May
|
Jun
|
Jul
|
Aug
|
Sep
|
Oct
|
Nov
|
Dec
|
From: Gerrit E.G. H. <Ger...@be...> - 2007-08-06 19:59:24
|
L.S., I've been thinking a bit about this binary/native file data stuff and had a idea I'd like to discuss. First off, I'm pretty sure by now that 'native binary format' for the storage is a Good Thing(tm) as it makes CRM act _fast_. (And that's precisely what I need.) Why is it 'good'? Because crm uses a very fast mechanism to access the files called memory mapping (mmap(3) or the Win32 equivalent). Having crm translate/convert those bytes to/from 'native format' would either slow down the memory mapped file access (each read/write would have to convert int/double formats when mmap is used) or we'd have to throw memory mapping out the window, i.e. load every file into memory, convert once on load and once on save (for learn) causing an increase in both startup and memory usage overhead. Both these methods, aimed at providing a 'portable binary file format', would have serious impact on crm performance IMHO. (And it just might make the code a bit less legible too, but that can be handled in other ways.) Then another thought struck me: while scanning through my (rather short) archive of the mailing list, I've seen a few questions about the 'portability' of these files across multiple CRM versions. (e.g. the 64-bit discussion with regards to a new Debian build, IIRC) I don't have a proper solution for resolving the latter, but we might consider ways to ensure crm binary formats are 'forwards portable' across platforms _and_ versions/releases. (forwards portable = anything before this moment may be 'broken', but from now on we can offer portability) The idea is this (it has been done in a minor way before within crm): - start each file with a versioned header (I'll come back to that later) - the rest of the file is as it is right now: the hashes and all the other goodies, stored in a native format (int, double, etc.) for _fast_ every access/use. The way to provide the forward portability would be through providing an export/import mechanism (already exists for a few formats: cssdump) which can be used to export binary files to a portable (preferably text?) format, which can then be imported into another crm: different version and/or platform. A bit of a snag there: I've found that the current way of calculating the hashes is not portable, i.e. does not deliver the exact same values on different platforms. That has to be fixed before 'cross-platform' would be feasable. (As I don't see a 'hash converter' happening: hashes are essentially a one-way trip, so there's no way to convert hashes that were calculated based on slightly different bit patterns.) I'm not sure fixing the strnhash() code so as to calculate the same hash bitpattern on every platform is quite enough to ensure cross-platform compatibility though: the data in the files is sometimes derived from the hashes: that would imply those 'derivation paths' would also have to be bit-compatible across platforms. (I'm not sure, have to check again there if we stored derived data next to config params and hash series.) Let's assume however that we've got the hashes/import/export thing covered. And even if we don't we would be well aware of the limitations then. That leaves that 'versioned header' I mentioned. The versioned header should contain enough information for an export/import function to operate correctly: a) import all acceptable data, or b) report the incompatibility and hence the need to 'recreate/relearn' the files. Especially (b) is important as that'd enable (automated) upgrades to properly interact with the users: one would then be able to select wether to commence with an 'incompatible' upgrade or delay it until a later, opportune moment. To make it all cross compatible, the header would need to include these bits of info: - actual crm version - the classifier used - the platform (little/big endian, 32/64/... bits, etc.) - data type (just to make sure we're not messing with the wrong sort of stuff) so the im/exporter can handle possible data content / format changes and/or data compatibility checks. The header has to be in a cross-platform portable format so that every crm toolset of there can recognize copied/moved files and act accordingly: report or process the data. Given the 'memory mapped' file access approach of crm, this would also mean that the header is fixed length (bytes). The software can then easily jump over this header to get at the data itself. Conclusion ----------- I suggest the introduction of a cross-platform binary format, fixed width header for all crm binary files (css et al) to provide (at least) forward compatibility. For the remainder, all these files remain as is: data will be stored in a native format for fast access. The binary format header will include these information items (at least): - the crm version used to create the file - the platform (integer size and format(endianess), floating point size and format, structure alignment, etc.) - the classifier used to create the file - the data content type (some classifiers use multiple files) - space for future expansion (this is a research tool too: allow folks to add their own stuff which may not fit the header items above) The approach includes the existence of an export/import tool to convert the data to/from a cross-platform portable text format, where applicable. At least, the im/exporter will then be able to accurately predict whether you can 'migrate' your data or need to start with a clean slate. This also depends on the sophistication of the im/exporter, of course, but the least benefit is a quick but accurate decision on 'migrateability' of your data. Which, I believe, doesn't exist now, unless you find 'better always rebuild from scratch' a good choice for every major/minor software upgrade. ;-) What are your thoughts on this matter? Is this worth persuing (and hence augmenting the code to support such a header from now on) or is this, well... Best regards, Ger |
From: <ste...@gm...> - 2007-07-22 21:55:08
|
-------- Original-Nachricht -------- Datum: Sun, 22 Jul 2007 23:40:28 +0200 Von: Paolo <oo...@us...> An: crm...@li... Betreff: Re: [Crm114-discuss] Problem with mailreaver.crm in 20070622-BlameTheInterns? > On Sun, Jul 22, 2007 at 08:57:14PM +0200, ste...@gm... wrote: > > Hello list > > you'd better post to crm114-general for bugs and 'general' use issues. > Okay. Will do so in the future. > > > > I have installed CRM114 20070622-BlameTheInterns and now when I try to > run the mailreaver I get the following error message: > > ->------<- > > Aw, crud. mailreaver.crm broke. Here's the error: :\_nl: > > crm: *WARNING* > > Too many close parenthesis in this math: 0.00 > 10.0) > > I'll try to keep working. > > This happened at line 545 of file mailreaver.crm > > known bug - check that mailreaver.crm @~545 looks like: > ... > # > # It's a pure classify. Three possibilities: > { > # Case 1 - It's spam. > eval /:@: :*:pr: < (0.0 - :*:thick_threshold:) :/ > goto /:looks_spam:/ > } > { > # Case 2 - It's good. > eval /:@: :*:pr: > :*:thick_threshold: :/ > goto /:looks_good:/ > } > ... > > there were/are unblanced '(' ')' in EVAL lines even in latest code from > src/. > Yes. I see it now. In Case 2. Thanks. > -- > paolo > Steve > ------------------------------------------------------------------------- > This SF.net email is sponsored by: Splunk Inc. > Still grepping through log files to find problems? Stop. > Now Search log events and configuration files using AJAX and a browser. > Download your FREE copy of Splunk now >> http://get.splunk.com/ > _______________________________________________ > Crm114-discuss mailing list > Crm...@li... > https://lists.sourceforge.net/lists/listinfo/crm114-discuss -- Psssst! Schon vom neuen GMX MultiMessenger gehört? Der kanns mit allen: http://www.gmx.net/de/go/multimessenger |
From: Paolo <oo...@us...> - 2007-07-22 21:40:40
|
On Sun, Jul 22, 2007 at 08:57:14PM +0200, ste...@gm... wrote: > Hello list you'd better post to crm114-general for bugs and 'general' use issues. > > I have installed CRM114 20070622-BlameTheInterns and now when I try to run the mailreaver I get the following error message: > ->------<- > Aw, crud. mailreaver.crm broke. Here's the error: :\_nl: > crm: *WARNING* > Too many close parenthesis in this math: 0.00 > 10.0) > I'll try to keep working. > This happened at line 545 of file mailreaver.crm known bug - check that mailreaver.crm @~545 looks like: ... # # It's a pure classify. Three possibilities: { # Case 1 - It's spam. eval /:@: :*:pr: < (0.0 - :*:thick_threshold:) :/ goto /:looks_spam:/ } { # Case 2 - It's good. eval /:@: :*:pr: > :*:thick_threshold: :/ goto /:looks_good:/ } ... there were/are unblanced '(' ')' in EVAL lines even in latest code from src/. -- paolo |
From: <ste...@gm...> - 2007-07-22 18:57:24
|
Hello list I have installed CRM114 20070622-BlameTheInterns and now when I try to run the mailreaver I get the following error message: ->------<- Aw, crud. mailreaver.crm broke. Here's the error: :\_nl: crm: *WARNING* Too many close parenthesis in this math: 0.00 > 10.0) I'll try to keep working. This happened at line 545 of file mailreaver.crm ERROR: mailreaver.crm broke. Here's the error\: ERROR: crm: *WARNING* Too many close parenthesis in this math: 0.00 > 10.0) I'll try to keep working. This happened at line 545 of file mailreaver.crm ->------<- The configuration I use is this one: :verbose_startup: /SET/ :do_base64: /yes/ :mime_decoder: /openssl base64 -d/ :cache_dupe_command: /\/bin\/ln/ :general_fails_to: // :rejected_mail_exit_code: /0/ :accepted_mail_exit_code: /0/ :unsure_mail_exit_code: /0/ :program_fault_exit_code: /1/ :add_headers: /yes/ :add_verbose_stats: /yes/ :add_mailtrainer_report: /yes/ :add_extra_stuff: /no/ :spam_flag_subject_string: /ADV:/ :good_flag_subject_string: // :unsure_flag_subject_string: /UNS:/ :confirm_flag_subject_string: /TCF:/ :rewrites_enabled: /yes/ :log_to_allmail.txt: /no/ :log_all_mail_to_file: // :text_cache: /reaver_cache/ :trainer_invoke_command: /.\/mailtrainer.crm/ :trainer_randomizer_command: /.\/shuffle.crm/ :log_rejections: /yes/ :log_rejections_to_file: // :inoculations_enabled: /no/ :decision_length: /16000/ :expand_urls: /no/ :url_fetch_cmd: /wget -T 30 -O - / :url_trim_cmd: / head -c 16000 / :clf: /osbf/ :thick_threshold: /10.0/ :lcr: /[[:graph:]]+/ :undo_interruptus: /no/ :automatic_training: /no/ Is this a bug or do I have configured something the wrong way (The configuration used to work with 20070428-BlameSpamConf)? // Steve -- GMX FreeMail: 1 GB Postfach, 5 E-Mail-Adressen, 10 Free SMS. Alle Infos und kostenlose Anmeldung: http://www.gmx.net/de/go/freemail |
From: Tony G. <to...@of...> - 2007-05-27 21:12:55
|
In my tests, I found that running through spamassassin (with razor tests) and keeping its headers did better than giving crm114 the raw headers since spamassasin generalized the whole open-relay-identification issue. But haven't done the same test with recent classifiers On 5/24/07, Paolo <oo...@us...> wrote: > On Fri, May 25, 2007 at 01:53:48AM +0200, martin f krafft wrote: > > |formail -kxFrom: -xSubject: | crm ??? > ... > > This got me thinking: do headers contain valuable information for > > yes > > > Would it not make sense to simply crop all headers other than > > Subject and then to train and classify based on the subject and body > > depends on your stuff: if you want to classify your mail, other than > spam/good, you *may* want to keep just to: cc: subject: and body, > since you're mostly interested in the actual body contents. > > But for spam/good extensive tests have shown that headers play a big role, > though how useful such role is, depends on classifier (eg ifile(1) didn't > like them). > I'm used to trash all ID headers and zero date fields though, as that's > pure noise. > > See attached awk script to get an(other) idea to play with. > (you may want to (try to) re-implement it in crm as an exercise ;) > if so pls post back :) ) > > > > -- > paolo > > > ------------------------------------------------------------------------- > This SF.net email is sponsored by DB2 Express > Download DB2 Express C - the FREE version of DB2 express and take > control of your XML. No limits. Just data. Click to get it now. > http://sourceforge.net/powerbar/db2/ > _______________________________________________ > Crm114-discuss mailing list > Crm...@li... > https://lists.sourceforge.net/lists/listinfo/crm114-discuss > > > -- -- Tony Godshall g |
From: Paolo <oo...@us...> - 2007-05-25 08:00:16
|
On Thu, May 24, 2007 at 08:31:13PM +0200, martin f krafft wrote: > I am looking for a tool which can be trained to estimate > probabilities of a text to belong to any number of categories. For ... > Can crm114 be scripted to support this kind of learning? I give it did you check the sample gatlingsort.crm in the distrib/tarball? that might be a starting point. -- paolo |
From: Paolo <oo...@us...> - 2007-05-25 07:57:23
|
On Thu, May 24, 2007 at 08:17:06PM +0200, martin f krafft wrote: > files. Specifically, I want my *.mfp and *.cf files in an etc-like > directory, while the *.css files and the reaver_cache directory in > a var-like directory. > The :datadir: variable defined in mailreaver.crm seems to affect seems it's gone in latest release ... > From the source, it looks as if *.css file and *.cf files are > accessed under :*:fileprefix: while *.mfp files are under the seems correct. You may want to get latest version, which allows you to set :var: in INSERT too. > Aren't *.mfp files configuration files (which are only changed by > the user) and should live next to mailfilter.cf, while *.css files > and the reaver_cache are var-like data? Would you consider a patch that depends on your situation - you may have any mix of system+user 's configs+vars. The default situation for the 'kit' is a 'per user' setup, so yes, *.cf, *.mfp are configs, *.crm are progs, *.css and *cache are vars; in Debian you'd likely have them in ~/.crm114/etc (and perhaps some system-default in /etc/crm114 and/or /etc/defaults/crm114), /usr/lib/crm114 and ~/.crm114/{cache,data} or /var/lib/crm114/<uid>/{cache,data} resp. or something the like. > no micro$oft components were used > in the creation or posting of this email. > therefore, it is 100% virus free > and does not use html by default (yuck!). sadly, that's not true. -- paolo |
From: Paolo <oo...@us...> - 2007-05-25 05:22:17
|
On Fri, May 25, 2007 at 01:53:48AM +0200, martin f krafft wrote: > |formail -kxFrom: -xSubject: | crm ??? ... > This got me thinking: do headers contain valuable information for yes > Would it not make sense to simply crop all headers other than > Subject and then to train and classify based on the subject and body depends on your stuff: if you want to classify your mail, other than spam/good, you *may* want to keep just to: cc: subject: and body, since you're mostly interested in the actual body contents. But for spam/good extensive tests have shown that headers play a big role, though how useful such role is, depends on classifier (eg ifile(1) didn't like them). I'm used to trash all ID headers and zero date fields though, as that's pure noise. See attached awk script to get an(other) idea to play with. (you may want to (try to) re-implement it in crm as an exercise ;) if so pls post back :) ) -- paolo |
From: Tony G. <to...@of...> - 2007-05-25 04:27:01
|
Good idea. Run the test and report back on the improvement, if any, in accuracy. On 5/24/07, martin f krafft <ma...@ma...> wrote: > Hi, > > a colleague showed me his way of invoking crm114, which is > > |formail -kxFrom: -xSubject: | crm =85 > > so effectively he deletes all headers but From and Subject. > > This got me thinking: do headers contain valuable information for > crm114? There is a lot of redundant or constant information in > there, and spam these days comes via random routes and from random > senders with random subjects anyway. > > Would it not make sense to simply crop all headers other than > Subject and then to train and classify based on the subject and body > only? > > Cheers, > > -- > martin; (greetings from the heart of the sun.) > \____ echo mailto: !#^."<*>"|tr "<*> mailto:" net@madduck > > spamtraps: mad...@ma... > > http://www.vcnet.com/bms/ > > -----BEGIN PGP SIGNATURE----- > Version: GnuPG v1.4.6 (GNU/Linux) > > iD8DBQFGViWMIgvIgzMMSnURAkm4AKDi89/b1kqAR7lBzUkZffFOiHj9twCfWzgZ > lVVf9sUgjl+HeQ1iR6bFR8M=3D > =3Doznd > -----END PGP SIGNATURE----- > > ------------------------------------------------------------------------- > This SF.net email is sponsored by DB2 Express > Download DB2 Express C - the FREE version of DB2 express and take > control of your XML. No limits. Just data. Click to get it now. > http://sourceforge.net/powerbar/db2/ > _______________________________________________ > Crm114-discuss mailing list > Crm...@li... > https://lists.sourceforge.net/lists/listinfo/crm114-discuss > > --=20 -- Tony Godshall g |
From: martin f k. <ma...@ma...> - 2007-05-24 23:53:54
|
Hi, a colleague showed me his way of invoking crm114, which is |formail -kxFrom: -xSubject: | crm =E2=80=A6 so effectively he deletes all headers but From and Subject. This got me thinking: do headers contain valuable information for crm114? There is a lot of redundant or constant information in there, and spam these days comes via random routes and from random senders with random subjects anyway. Would it not make sense to simply crop all headers other than Subject and then to train and classify based on the subject and body only? Cheers, --=20 martin; (greetings from the heart of the sun.) \____ echo mailto: !#^."<*>"|tr "<*> mailto:" net@madduck =20 spamtraps: mad...@ma... =20 http://www.vcnet.com/bms/ |
From: martin f k. <ma...@ma...> - 2007-05-24 18:31:19
|
Hi there, I am looking for a tool which can be trained to estimate probabilities of a text to belong to any number of categories. For instance, a given text may have a 65% resemblance of texts in category/bucket A, 90% resemblance to texts in category/bucket B, and 5% resemblance to texts in category/bucket C. Can crm114 be scripted to support this kind of learning? I give it a bunch of texts each with a set of categories and it learns? What I am trying to create is a learning, statistics-based tagging system for email. Folders are not good enough for my mail workflow as messages can apply to multiple contexts. Gmail's labels and e.g. evolutions virtual folders are more like it, but I would much rather learn my labels than set rules for them. Can it be done? --=20 martin; (greetings from the heart of the sun.) \____ echo mailto: !#^."<*>"|tr "<*> mailto:" net@madduck =20 spamtraps: mad...@ma... =20 "to have the reputation of possessing the most perfect social tact, talk to every woman as if you loved her, and to every man as if he bored you." -- oscar wilde |
From: martin f k. <ma...@ma...> - 2007-05-24 18:17:25
|
Hi, I am now a happy crm114 (mailreaver) user who would like to become even more happy by being able to separate variable and configuration files. Specifically, I want my *.mfp and *.cf files in an etc-like directory, while the *.css files and the reaver_cache directory in a var-like directory. I call mailreaver like so: /usr/share/crm114/mailreaver.crm -u $HOME/.etc/crm114/ and this works, except that it expects the *.css files in $HOME/.etc/crm114/ as well, and it will create the reaver_cache there too. I can fix the reaver_cache location in mailfilter.cf (simply by setting :text_cache: to a relative path of the desired cache location), but I cannot get the *.css files out of $HOME/.etc/crm114. The :datadir: variable defined in mailreaver.crm seems to affect only the *.mfp files, and calling mailreaver.crm from the data dir also uses the *.css files in $HOME/.etc/crm114. =46rom the source, it looks as if *.css file and *.cf files are accessed under :*:fileprefix: while *.mfp files are under the datadir? Aren't *.mfp files configuration files (which are only changed by the user) and should live next to mailfilter.cf, while *.css files and the reaver_cache are var-like data? Would you consider a patch to reflect this? --=20 martin; (greetings from the heart of the sun.) \____ echo mailto: !#^."<*>"|tr "<*> mailto:" net@madduck =20 spamtraps: mad...@ma... =20 no micro$oft components were used in the creation or posting of this email. therefore, it is 100% virus free and does not use html by default (yuck!). |
From: <li...@wc...> - 2007-02-25 21:28:44
|
Hey all, CRM114 newb here. Originally looked at CRM114 somewhere around BlameEasterBunny and was scared away. I've recently come into the need for a non-email text classifier and delved back into the world of CRM114 - and very glad I did. This is heady stuff and my mind is swimming with the possibilities. 'Twould be a shame to shackle such power to simple email filtering, IMHO. Even though spam filtering appears by far to be the most common use, I was disgusted to find CRM114 in the Mail ports on FreeBSD instead of Textproc. :) After wading through "CRM114 Revealed" I happily noticed the mention that some of the classifiers support N-way choices. Prior to reading this, I assumed all classifiers were two-way. I envisioned a tree that would allow for N classifications by doing up to hundreds of binary classifications in series. It appears like that method is unnecessary. Here's my question: which of the seven classifiers included with CRM114 BlameDalkey support N-way? I've made some guesses, but a lot of this stuff is yet a ways above me. BTW, this info would be a nice addition to the classifier summary chart on page 171 on the book. Happy CRMming! Brad Waite |
From: martin f k. <ma...@ma...> - 2006-10-06 10:37:50
|
I am aware that crm114 uses char, not wchar. However, I wonder whether and how this affects its applicability to filter email. From what I understand, mail messages are 8 (or even 7)-bit clean, so if a sender uses Unicode, crm114 will actually just see the bytes making up the wide characters, and learn those. If it previously learnt to categorise mail about increasing your male member's dimensions as spam, it won't be able to apply that knowledge to unicode mail on the same topic. Are there any other limitations of which to be aware? Cheers, --=20 martin; (greetings from the heart of the sun.) \____ echo mailto: !#^."<*>"|tr "<*> mailto:" net@madduck =20 spamtraps: mad...@ma... =20 "out of the crooked timber of humanity, no straight thing was ever made." -- imanuel kant |
From: martin f k. <ma...@ma...> - 2006-10-06 10:37:48
|
Dear list, I am in search for a new way to handle mail without deviating from my current use of offlineimap and mutt. What I want are tags on messages, rather than the classic storing mail into single locations in a folder hierarchy. The reason is simple: a mail may be relevant in different contexts, and copying, hardlinking, or symlinking messages around isn't something with which mutt can deal. So I want virtual folders (where a folder is simply a virtual container of all messages that have a certain tag; thus, I am equating virtual folders with tags). I want to use an advanced learning technique to assign tags to messages, and it seems natural to come here. In the words of Dr. Strangelove: "based on the findings of the report, my conclusion was that this idea was not a practical deterrent for reasons which at this moment must be all too obvious." I found the crm114 book and I intend to work through it in my copious amounts of spare time (haha!), but to get me jumpstarted, I am looking for examples which use crm114 in a multibucket type way. I imagine that I'd predefine the categories and then train the filter to assign them. I don't want N buckets and have crm114 assign each message to a single bucket of those N, but I'd want crm114 to say "this message applies to buckets 2, 3, 5, and 9". Is this possible? I would think so, but I can't wrap my head around two issues: - instead of running a separate programme to decide yes/no for each tag, how would I go about assigning tags with a single programme? Does anyone have a crm114 programme that does something along those lines? Thanks a lot for your time! --=20 martin; (greetings from the heart of the sun.) \____ echo mailto: !#^."<*>"|tr "<*> mailto:" net@madduck =20 spamtraps: mad...@ma... =20 quidquid latine dictum sit, altum viditur. |
From: <fl...@ea...> - 2006-08-23 06:59:52
|
hi how crm114 detect image based spam ? thanks wong |
From: glaucius j. <gla...@gm...> - 2006-06-12 18:01:25
|
Hi everyone I'm using postfix as a MTA, but my real mailserver is an exchange 2003, my postfix acts as a relay/gateway, can I use crm114 with postfix as a gateway/relay ??? tanks best regards !! |
From: <van...@gl...> - 2006-05-18 09:14:22
|
<http://www.medifestindia.com/images/logo-with-title_02.jpg> <http://www.medifestindia.com/images/header-pic_03.jpg> <http://www.medifestindia.com/images/glob.gif> Deal with 1/3rd of the world in a single shot, focus your goals for the upcoming biggest ever Mega Healthcare Festival in the INDIAN capital, New Delhi. -- Complete Medical Healthcare Industry under one roof -- <http://www.medifestindia.com/images/pic-1_06.jpg> <http://www.medifestindia.com/images/whyindia_05.gif> Indian Government has India <http://www.medifestindia.com/images/doted.gif> Laid high budgetary outlay for healthcare <http://www.medifestindia.com/images/doted.gif> Reduced import custom duties and simplified import procedures <http://www.medifestindia.com/images/doted.gif> Declared Tax exemption and financial incentives for new hospitals and healthcare entrepreneurs <http://www.medifestindia.com/images/doted.gif> Planned Bulk purchases for upgrading the health infrastructure <http://www.medifestindia.com/images/doted.gif> Allowed Foreign Direct Investment upto 51% in hospital services & upto 26% in health-insurance <http://www.medifestindia.com/images/doted.gif> Privatised the Medical Insurance sector <http://www.medifestindia.com/images/doted.gif> Imports the maximum of medical equipment (approx.85%) <http://www.medifestindia.com/images/doted.gif> Has huge number of upcoming hospitals and healthcare centres <http://www.medifestindia.com/images/doted.gif> Is becoming the most popular healthcare tourism destination on the planet. <http://www.medifestindia.com/images/doted.gif> Is the second largest populated country and incidence of diseases is very high <http://www.medifestindia.com/images/pic-2_08.jpg> India's healthcare industry grew at 13% per annum over the last decade and is currently growing at 17% annually. <http://www.medifestindia.com/images/project.gif> 2003 - Rs 95,000 Crores (US $ 19.25 Billion) 2008 - Rs. 1,75,000 Crores (US $ 40 Billion) 2012 - Rs. 3,20,000 Crores (US $ 70 Billion) Visitors Decision makers from : Medical and Para medical fraternity Health Organisations Hospitals Manufacturing Houses Research & Development Organisations Government Agencies Exhibitors Worldwide Exhibitors of : Medical Technology Pharmaceuticals Healthcare Informatics Medical Consumables and disposables Diagnostic Kits & reagents Healthcare Products And other related For more information please fill the form below Company: Contact Person: E-mail: Country: US States Afghanistan Albania Algeria Andorra Angola Argentina Armenia Australia Austria Azerbaijan Bahamas Bahrain Bangladesh Barbados Belarus Belgium Belize Benin Bhutan Bolivia Botswana Brazil Brunei Bulgaria Burkina_Faso Burundi Cambodia Cameroon Canada Cape_Verde Chad Chile China Colombia Comoros Costa_Rica Croatia Cuba Cyprus Czech_Republic Cote_d_Ivoire Denmark Djibouti Dominica Dominican_Republic Ecuador Egypt El_Salvador Equatorial_Guinea Eritrea Estonia Ethiopia Fiji Finland France Gambia Georgia Germany Ghana Greece Grenada Guatemala Guinea Guinea_Bissau Guyana Haiti Honduras Hungary Iceland India Indonesia Iran Iraq Ireland Israel Italy Jamaica Japan Jordan Kazakhstan Kenya Kiribati Korea__North Korea__South Kuwait Kyrgyzstan Laos Latvia Lebanon Lesotho Liberia Libya Liechtenstein Lithuania Luxembourg Madagascar Malawi Malaysia Maldives Mali Malta Marshall_Islands Mauritania Mauritius Mexico Moldova Monaco Mongolia Morocco Mozambique Myanmar Namibia Nauru Nepal Nicaragua Niger Nigeria Norway Oman Pakistan Palau Papua_New_Guinea Paraguay Peru Philippines Poland Portugal Qatar Romania Russia Rwanda Saint_Lucia Samoa San_Marino Saudi_Arabia Senegal Seychelles Sierra_Leone Singapore Slovakia Slovenia Solomon_Islands South_Africa Spain Sri_Lanka Sudan Suriname Swaziland Sweden Switzerland Syria Taiwan Tajikistan Tanzania Thailand Togo Tonga Tunisia Turkey Turkmenistan Tuvalu Uganda Ukraine United_Kingdom United_States Uruguay Uzbekistan Vanuatu Vatican_City Venezuela Vietnam Western_Sahara Yemen Zambia Zimbabwe Products: Website: <http://www.medifestindia.com/images/btn_tell-a-friend_10.gif> Unsubscribe <http://www.medifestindia.com/Thanks.aspx?UNB={1}> |
From: Matt K. <mat...@gm...> - 2005-10-27 18:14:32
|
I think it would be the cat's pajamas to feed a pile of unclassified documents into the hyperspace classifier and get back a matrix of the 'distance' between each pair of docs. This does not seem to be possible fro= m within CRM114, at least from my reading of The Book. (CRM114 Revealed 2005-09-29). Am I mistaken, or do I need to get into the guts of CRM itself to get this feature? -- b++ pr(a+=3Db) while (pr(b+=3Da) < 10000) |
From: Raul M. <mo...@ma...> - 2005-09-26 13:07:08
|
On Mon, Sep 26, 2005 at 02:47:56PM +0400, =E7=C1=CC=C9=CE=C1 =E9=CF=C6=C9= =CE=C1 wrote: > I wish to know very much how it works from inside- all mathematical > points and if it is possible the proofs of its working (again not from > practical point, but theoretical). The development of the CRM classification algorithms seem to be more based on science than on axioms. That said, depending on how it's configured, it can be a markovian filter= , a baysian filter, or something that people maybe haven't developed a mathematical basis yet to describe (for example, the hyperspace filter). It's still under development (and might remain that way until spammers stop developing new spam techniques)... But to have a rigorous mathematical basis for its development, we'd need a rigorous mathematical model for what people consider to be spam. But different people have different ideas of such things, so... We rely on Bill's excellent taste in algorithms (and the dedication of the CRM developers to testing) over more abstracted approaches. That said, if you want to develop theories of probability to account for how well it works, I'm sure many of us here would be interested in reading them. --=20 Raul |
From: <ga...@ma...> - 2005-09-26 10:48:08
|
Hello everybody. I am a student of mathimatical department of MIPT. Just recently I learnt about the existance of your CRM114 algorithm, and was very much interested in it. But I found very little information about its entity. I wish to know very much how it works from inside- all mathematical points and if it is possible the proofs of its working (again not from practical point, but theoretical). If it is not a secret, please send me a reference on it in the internet or an article. But I want to specify that I need knowlage nit in CRM-language, but in algorithm. Thank you very much. Sencirely Galina. |
From: <wed...@co...> - 2005-03-30 19:39:25
|
<!-- This plain text message area is for recipients who have email programs that= do not support html emails (almost every modern email client software progra= m can accept html emails). If you want to add an alternate plain text messag= e for those who cannot receive html emails, enter it here. //--> <HTML> <HEAD> <META NAME=3D"GENERATOR" Content=3D"Microsoft DHTML Editing Control"> <TITLE></TITLE> </HEAD> <BODY> <P class=3DMsoNormal style=3D"MARGIN: 0in 0in 0pt">We have some wholesale Sa= fe &=20 Natural pet products you may be interested in adding to your inventory.</P> <P class=3DMsoNormal style=3D"MARGIN: 0in 0in 0pt; tab-stops: 155.25pt"><SPA= N=20 style=3D"mso-tab-count: 1"> = &nbs= p; &n= bsp; = =20 </SPAN></P> <P class=3DMsoNormal style=3D"MARGIN: 0in 0in 0pt">Great Deals! <A=20 href=3D"http://www.wedothat.org/"><SPAN=20 style=3D"COLOR: windowtext; TEXT-DECORATION: none; text-underline: none">ww= w.wedothat.org</SPAN></A></P> <P class=3DMsoNormal style=3D"MARGIN: 0in 0in 0pt"> <?xml:namespace pre= fix =3D o=20 ns =3D "urn:schemas-microsoft-com:office:office" /><o:p></o:p></P> <P class=3DMsoNormal style=3D"MARGIN: 0in 0in 0pt"> <o:p></o:p></P> <P class=3DMsoNormal style=3D"MARGIN: 0in 0in 0pt">Thank you for your time.<= /P> <P class=3DMsoNormal style=3D"MARGIN: 0in 0in 0pt">Jackie Gallipo</P> <P class=3DMsoNormal style=3D"MARGIN: 0in 0in 0pt">We Do That</P> <P class=3DMsoNormal style=3D"MARGIN: 0in 0in 0pt">360-332-1966</P> <P class=3DMsoNormal style=3D"MARGIN: 0in 0in 0pt"><A=20 href=3D"http://www.wedothat.org/"><SPAN=20 style=3D"COLOR: windowtext; TEXT-DECORATION: none; text-underline: none">ww= w.wedothat.org</SPAN></A></P> <P class=3DMsoNormal style=3D"MARGIN: 0in 0in 0pt"><A=20 href=3D"mailto:wed...@co..."><SPAN=20 style=3D"COLOR: windowtext; TEXT-DECORATION: none; text-underline: none">we= do...@co...</SPAN></A></P> <P class=3DMsoNormal style=3D"MARGIN: 0in 0in 0pt"> </P> <P class=3DMsoNormal style=3D"MARGIN: 0in 0in 0pt">If you want to unsubscrib= e to=20 this e-mail, please <A href=3D"mailto:wed...@co...">click here</A><= /P> <P class=3DMsoNormal style=3D"MARGIN: 0in 0in 0pt"><U> <o:p></o:p></U><= /P> <P> </P> </BODY> </HTML> |
From: Sami D. <sk...@fr...> - 2005-01-28 21:08:13
|
Hi, I'm currently writing a SBPH/Markovian Mail Filter in C# (https://savannah.nongnu.org/projects/mailprobe/). However, I'm wondering about your implementation since I'm partly (mostly ?) taking ideas from your project. Here's my problem : Let say that Bayes formula is : P(in class / feature) = [P(feat / in class) * P( in class) ] / [P(feat/class) * P(in class) + P(feat/ not class) * P (not class)] Using Markovian Matching, we get local probabilities of the form : Plocal-spam = 0.5 + [(Nspam - NnotSpam)*Weigt] / [C1 * (Nspam + Nnonspam + C2) * WeightMax] => do you actually use this formula ? BTW, does anyone know where this formula comes from ? First question : 1) Plocal-spam is equivalent to the P(feat / class) part in the Bayesian Rule, right ? 2) I'd love if you could help me understand the following part of your classify_details.txt : "We start assuming that the chance of spam is 50/50" => This means we set P(in class) = 0.5, right ? (in the baysian rule) and, for N classes, we start the chance of a class at 1/N ? "We count up the total number of features in the good versus evil feature .css files. We use these counts to normalize the chances of good versus evil features, so if your training sets are mostly good, it doesnt predispose the filter to think that everything is good." => Let's say the total feature count in the "good" class is 10 and the total feature count in the "evil" class is 20. This means we get 30 features totally. What do you call exactly "normalize" ? In Bayes Formula, here's what we need to calculate : -a) P(feat/in class) -b) P(in class) -c) P(feat / not class) -d) P( not class) What do you need to "normalize" exactly ? P(feat / in class) and P(feat / not class) => doesn't depend on the other classes. If a specific feature was counted 4 times in the class c, and class c has 100 features, then P(feat / in class) = 4/100 if we calculate the P(feat / in class) by using the formula : Plocal-spam = 0.5 + [(Nspam - NnotSpam)*Weigt] / [C1 * (Nspam + Nnonspam + C2) * WeightMax] then does "normalizing" mean that .. total count for "evil": 20 total count for "good" : 10 => then, if the database says that Nspam = 2 and NnotSpam = 3 (for a certain feature), then do we say : Nspam = 2 and Nnotspam = (20/10) *3 = 6 and then calculate the probability ? P( in class) and P(not class) => didn't we just say it was 0.5 ? "We repeated form a feature with the polynomials, check the .css files to see what the counts of that feature are for spam and nonspam, and use the counts to calculate P(A/S) and P(A/NS) [ remember, we correct for the fact that we may have different total counts in the spam and nonspam categories ] Same question as before. what about normalization ? "We also bound P(A/S) and P(A/NS) to prevent any 0.0 or 1.0 probabilities from saturating the system" => yes, sure, ok. I still haven't understood the [1/featurecount+2, 1 - 1/featurecount+2] limit.. where does it come from ? empirical ? and what about "and then to add further uncertainty to that bound additionally buy a factor of 1/(featurecount+1) => do you use random number or something like that ? "Once we have P(A/S) and P(A/NS), we calculate the new P(S) and P(NS). Then we get the next feature out of the polynomial hash pipeline and repeat until we hit the end of the text".. Ok, so you mean that you set P(S) and P(NS) to the new value, and then recalculate the Bayesian formula for the next feature. I'm just wondering why everybody seems to use a different way of doing things : usually, people seem to use something like - supress the denominator in the Bayesian Rule (supposed to be a constant ? Doesn't seem to be one in your formula - then to calculate P(Features / class) * P(Class), they usually calculate : P(Feature1/class) * P(feature2/class)*....*P(featureN/class) *P(class) this way of calculating and yours are different, so I'm really wondering which one is the best ;p 3) About TOE: I'm wondering about what you call TOE: Do you mean that the user only has to train the filter on errors, or that CRM114 keeps re-learning error-prone emails until they are classified correctly ? I think that's all. Thank you very much for your help. Sami Dalouche |
From: kent s. <ke...@ya...> - 2004-11-11 23:23:27
|
sorry for off-topic here. I am looking for tre-python.c The tre 0.7.0 is suppose to come with a python binding. However, when cd into the distributed python directory and python setup.py install: gcc: tre-python.c: No such file or directory gcc: no input files Can somebody help? Thank you. |
From: Chris M. <ch...@im...> - 2004-08-03 22:23:43
|
Hi there. I've been using CRM114 quite successfully for a number of months. A day or two ago, it simply stopped working. I haven't made any changes to the system that I'm aware of. According to the message headers, CRM114 doesn't seem to be getting the messages anymore, but I'm not sure why. Details of my config: Redhat enterprise server 3.1 Postfix 2.0.16 Procmail delivery my procmailrc: :0fw: .msgid.lock | /usr/bin/crm -u /home/chrism mailfilter.crm :0: * ^X-CRM114-Status: SPAM.* mail/crm-spam This worked flawlessly and then *boom*....like someone turned off the light switch. Any suggestions would be very very much appreciated. Best regards, Chris |