Update of /cvsroot/popfile/engine
In directory sc8-pr-cvs1:/tmp/cvs-serv17259
Modified Files:
traintest.pl
Log Message:
add stopword toggle (-stopwords 0 to test without stopwords)
Index: traintest.pl
===================================================================
RCS file: /cvsroot/popfile/engine/traintest.pl,v
retrieving revision 1.2
retrieving revision 1.3
diff -C2 -d -r1.2 -r1.3
*** traintest.pl 16 Mar 2003 00:57:16 -0000 1.2
--- traintest.pl 16 Mar 2003 01:32:31 -0000 1.3
***************
*** 16,19 ****
--- 16,20 ----
my $DEFAULT_CSV = 'auto';
my $DEFAULT_TOE = 1;
+ my $DEFAULT_STOP = 1;
my $DEFAULT_WINDOW1 = 100;
my $DEFAULT_WINDOW2 = 500;
***************
*** 64,67 ****
--- 65,69 ----
$config->parameter("csv",$DEFAULT_CSV);
$config->parameter("toe",$DEFAULT_TOE);
+ $config->parameter("stopwords",$DEFAULT_STOP);
$config->parameter("window1",$DEFAULT_WINDOW1);
$config->parameter("window2",$DEFAULT_WINDOW2);
***************
*** 87,90 ****
--- 89,101 ----
}
+ if ( $config->parameter('stopwords') ne $DEFAULT_STOP ) {
+ $file .= "_";
+ if ($config->parameter('stopwords') != 1) {
+ $file .= "no";
+ }
+ $file .= "stop";
+ }
+
+
$file .= ".csv";
}
***************
*** 251,254 ****
--- 262,271 ----
$b->{unclassified__} = ($c->parameter("bayes_unclassified_probability") || 0.5);
+ # test with or without stop-words
+ if ( $c->parameter("stopwords") eq 0 ) {
+ $b->{parser__}->{mangle__}->{stop__} = {};
+ $b->{mangler__}->{stop__} = {};
+ }
+
my $archive = $c->parameter("ui_archive_dir");
***************
*** 398,402 ****
print " -archive_dir: Location to seek an archive\n";
print " -csv: Filename to save CSV log to, \"auto\" generates a filename\n";
! print " -toe: Train Only Errors, defaults to $DEFAULT_TOE";
}
--- 415,420 ----
print " -archive_dir: Location to seek an archive\n";
print " -csv: Filename to save CSV log to, \"auto\" generates a filename\n";
! print " -toe: Train Only Errors, defaults to $DEFAULT_TOE\n";
! print " -stopwords: Use stop-words, defaults to $DEFAULT_STOP\n";
}
|
