dbacl - digramic Bayesian classifier Git
commandline multiclass email and text filter
Brought to you by:
lbreyer
Menu
▾
▴
[0ec8fa]: / doc / example3.risk Maximize Restore History
17 lines (15 with data), 489 Bytes
# Same as example1.risk but with a small increase in two/three,
# and an regular expression dependent loss matrix
categories {
one, two, three
}
prior {
2, 1, 1
}
loss_matrix {
"^proportion: ([0-9.]+)" one [ 0, (1+$1)^complexity, (1.2)^complexity ]
"" one [ 0, (1.1)^complexity, (1.1)^complexity ]
"Polly" two [inf, 0, inf ]
"" two [(1.1)^complexity, 0, (2.0)^complexity ]
"" three [(1.5)^complexity, (1.01)^complexity, 0 ]
}