Amis - A maximum entropy estimator for feature forests

Amis is a parameter estimator for maximun entropy models [1]. Given a set of events as training data, the program outputs parameters that maximize the likelihood of the training data. The software supports the following functions.

• Parameter estimation algorithm for feature forests [2]
• Parameter estimation algorithms GIS, IIS [3], and limited-memory BFGS [4]
• Selectable features: binary, integer, and real
• Compression of event files

Download

• amis-2.0.tar.gz
• amis-2.95.tar.gz (Support for gcc3)
• amis-3.0.tar.gz
• amis-3.1.tar.gz (Support for gcc3.4)
• amis-3.1.1.tar.gz

Quick Installation

% tar xvzf amis-x.y.z.tar.gz
% cd amis-x.y.z
% ./configure
% make install

This procedure installs "/usr/local/bin/amis" (you must have an access to "/usr/local/bin/"). If you want to install it in another directory, see Amis manual.

Manuals

• Amis manual
• Amis manual (Japanese)

References

[1] Adam L. Berger, Stephen A. Della Pietra and Vincent J. Della Pietra. A maximum entropy approach to natural language processing. Computational Linguistics, 22(1):39-71, 1996.

[2] Yusuke Miyao and Jun'ichi Tsujii. Maximum entropy estimation for feature forests. In Proc. HLT2002.

[3] Stephen A. Della Pietra, Vincent J. Della Pietra and John Lafferty. Inducing features of random fields. IEEE Transactions on Pattern Analysis and Machine Intelligence, 19(4):380-393, 1997.

[4] Jorge Nocedal. Updating quasi-Newton matrices with limited storage. Mathematics of Computation, 35:773-783, 1980.