MSTParser

alpha

MSTParser is a non-projective dependency parser that searches for maxi

5.0 Stars (3)
21 Downloads (This Week)
Last Update:
Download mstparser-0.5.1-src.tgz
Browse All Files
Windows Mac Linux

Description

MSTParser is a non-projective dependency parser that searches for maximum spanning trees over directed graphs. Models of dependency structure are based on large-margin discriminative training methods. Projective parsing is also supported.

mstparser 0.5.1 is now available via Maven Central. If you use Maven as your build tool, then you can add it as a dependency in your pom.xml file:

<dependency>
<groupId>net.sourceforge.mstparser</groupId>
<artifactId>mstparser</artifactId>
<version>0.5.1</version>
</dependency>

MSTParser Web Site

Features

  • First and second order projective and non projective parsing
  • Perceptron and k-best MIRA training
  • Edge-wise confidence estimation (as of v0.5.0)

Update Notifications





User Ratings

★★★★★
★★★★
★★★
★★
3
0
0
0
0
ease 1 of 5 2 of 5 3 of 5 4 of 5 5 of 5 0 / 5
features 1 of 5 2 of 5 3 of 5 4 of 5 5 of 5 0 / 5
design 1 of 5 2 of 5 3 of 5 4 of 5 5 of 5 0 / 5
support 1 of 5 2 of 5 3 of 5 4 of 5 5 of 5 0 / 5
Write a Review

User Reviews

  • mercury3
    1 of 5 2 of 5 3 of 5 4 of 5 5 of 5

    You can give me how to run MST Parser to generate k-best trees for one sentence (not the best tree as default). Many thanks!

    Posted 04/21/2013
  • markbogner
    1 of 5 2 of 5 3 of 5 4 of 5 5 of 5

    Works well for me. Will use it again in the future.

    Posted 11/17/2012
  • charles-young
    1 of 5 2 of 5 3 of 5 4 of 5 5 of 5

    A great application that is true to its promise. Thank you very much.

    Posted 11/16/2012
Read more reviews

Additional Project Details

Intended Audience

Developers, Information Technology, Science/Research

User Interface

Command-line

Programming Language

Java, Python

Registered

2006-09-28
Screenshots can attract more users to your project.
Features can attract more users to your project.

Icons must be PNG, GIF, or JPEG and less than 1 MiB in size. They will be displayed as 48x48 images.