Lingua-EN-SENNA
view release on metacpan or search on metacpan
lib/Lingua/EN/SENNA/third-party/senna/doc/index.html view on Meta::CPAN
<!DOCTYPE html PUBLIC "-//W3C//DTD XHTML 1.0 Transitional//EN" "http://www.w3.org/TR/xhtml1/DTD/xhtml1-transitional.dtd">
<html xmlns="http://www.w3.org/1999/xhtml" xml:lang="en" lang="en">
<head>
<meta http-equiv="Content-Type" content="text/html; charset=UTF-8" />
<title>
SENNA
</title>
<link href="senna.css" rel="stylesheet" type="text/css" />
</head>
<body>
<center>
<img src="senna.jpg">
</center>
<h1>SENNA</h1>
<div class="level1">
<p>
SENNA is a software distributed under
a <a href="license.html">non-commercial license</a>, which outputs a host of
Natural Language Processing (NLP) predictions: part-of-speech (POS) tags,
chunking (CHK), name entity recognition (NER), semantic role labeling
(SRL) and syntactic parsing (PSG).
</p>
<p>
SENNA is fast because it uses a simple architecture, self-contained
because it does not rely on the output of existing NLP system, and accurate
because it offers state-of-the-art or near state-of-the-art performance.
</p>
<p>
SENNA is written in ANSI C, with about 3500 lines of code. It requires
about 200MB of RAM and should run on any IEEE floating point computer.
</p>
<p>
Proceed to the <a href="download.html">download page</a>. Read
the <a href="#compilation">compilation section</a> in you want to compile
SENNA yourself. Try out a <a href="#check">sanity check</a>. And read about
the <a href="#usage">usage</a>.
</p>
</div>
<h2>New in SENNA v3.0 (August 2011)</h2>
<div class="level2">
<p>
Here are the main changes compared to SENNA v2.0:
<ul>
<li><b>Syntactic parsing.</b></li>
<li>We now include our <b>original word embeddings</b>, used to trained each task.</li>
<li>Bug correction: now outputs correctly tokens made of numbers (instead of replacing numbers by "0").</li>
<li>Option <code>-offsettags</code>, which outputs start/end offsets (in the sentence) of each token.</li>
</ul>
</p>
<p>
DISCLAIMER: Our word embeddings differ
from <a href="http://metaoptimize.com/projects/wordreprs">Joseph Turian's
( run in 0.775 second using v1.01-cache-2.11-cpan-acf6aa7dc9e )