usage: spamfilter.py [-h] [--ham HAM] [--spam SPAM] [--mail MAIL] [-o OUTPUT] [-m MODEL] [-a {mi-logistic,mi-svm}] [-l LIMIT] [-d DICTLIMIT] [-r RAMLIMIT] {model,test}

Create a spamfilter model or apply a created filter. Create a model:

python spamfilter.py --ham ham.p --spam spam.p -o model.p model -a mi-logistic -l 5000 -d 5000

Test a model:

python spamfilter.py --mail mail.mbox -m model.p test -o tagged.mbox -l 10000

If the testing cmd takes MBOX as source then it separates the messages into two MBOXes. If the output is provided, then it creates tagged MBOX.

positional arguments: {model,test} Set program mode

optional arguments: -h, --help show this help message and exit --ham HAM A file containing HAM. --spam SPAM A file containing SPAM --mail MAIL A file containing unsorted mail -o OUTPUT, --output OUTPUT output file -m MODEL, --model MODEL model file -a {mi-logistic,mi-svm}, --algo {mi-logistic,mi-svm} Select training algorithms. -l LIMIT, --limit LIMIT Import set limit - memory limitation -d DICTLIMIT, --dictLimit DICTLIMIT Max dictionary size -r RAMLIMIT, --ramLimit RAMLIMIT Set RAM limit (MB) -- OOM prevent.

Tento projekt byl podpořen studentským fakultním grantem (SFG) Matematicko-fyzikální fakulty Univerzity Karlovy.