1.
Testing Paul Graham's Refinements
May 21, 2003 ... Paul Graham's Refinements to Bayesian Filtering ... Finally the r?.sp files were classified with that spam cutoff value, and the number of ...
2.
Greg's Bogofilter Page
Mar 24, 2004 ... 58 www.bgl.nu 0.184117 0.000080 0.000436 -0.00044 -7.73901 + 59 finger 0.187799 0.002326 0.012236 -0.01231 -4.40339 + 60 bgl.nu 0.123972 ...
3.
Bogofilter Training
Apr 18, 2003 ... Does full training or on-error training lead to fewer ... For this experiment I used a training corpus of 6372 nonspams and 6372 spam emails ...
4.
Testing Paul Graham's Refinements
May 21, 2003 ... Bogofilter is a Bayesian spam filter; background is presented on my general bogofilter page. Bogofilter runs a modification of Paul Graham's ...
5.
Creating executable R scripts
While this is handy, it seemed to me that it would be nice to be able to write executable R scripts capable of taking command-line parameters, as with sh, ...
6.
Bogofilter Calculations
Apr 12, 2003 ... cutoffs for beta will be set to match. # now in R: read.table("F/ns-0") -> ns0 read.table("F/ns-1") -> ns1 read.table("F/ns-2") -> ns2 ...
7.
Creating executable R scripts
The Bourne shell script presented below, named setR, takes advantage of R_PROFILE and .First to provide a #! equivalent for R scripts. ...
8.
Bogofilter Calculations
Apr 12, 2003 ... now in R: read.table("F/ns-0") -> ns0 read.table("F/ns-1") -> ns1 ... 6,6,6,7,7,7,8,8,8,9,9,9),sep=""),rep("r10",3)) beta ...
9.
Greg's Bogofilter Page
Mar 24, 2004 ... Bogofilter is a Bayesian spam filter, originally written by Eric S. Raymond with the purpose of creating a compact, fast implementation of ...
10.
Greg's Bogofilter Page
Mar 24, 2004 ... 58 www.bgl.nu 0.184117 0.000080 0.000436 -0.00044 -7.73901 + 59 finger 0.187799 0.002326 0.012236 -0.01231 -4.40339 + 60 bgl.nu 0.123972 ...
