You cannot select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.

16 lines
1.0 KiB
Plaintext

Library requirements for building Stupidfilter:
flex
libboost-serialization-dev
On Ubuntu 8.04, you should be able to just run
sudo apt-get install build-essential flex libboost-serialization-dev
make
make install
The Makefile may require a little bit of tweaking, specifically the path to your Boost serialization library, which in the test environment was /usr/lib/libboost_serialization-gcc42-mt-1_34_1.a
To run the StupidFilter directly just type bin/stupidfilter data/c_rbf
It will take data from standard in followed by a EOF and return a 0.000000 classification for stupid text and a 1.000000 for nonstupid text. Once we have regression working more accurately, this number will be actually be a floating point that describes how sure we are of the classification, so it's worth keeping it as a float in your implementations.
We have provided an example bash implementation in classify.sh. Note that we're normalizing whitespace with a call to sed. It's a good idea to strip HTML and normalize whitespace to avoid false positives.