Ecosyste.ms: Awesome
An open API service indexing awesome lists of open source software.
https://github.com/touhi99/statistical_dependency_parsing
SDP Lab Project - Arc-Eager transition-based dependency parsing with Averaged perceptron and extended features
https://github.com/touhi99/statistical_dependency_parsing
averaged-perceptron conll-2003 dependency-parsing parse-trees parsing perceptron transition
Last synced: about 6 hours ago
JSON representation
SDP Lab Project - Arc-Eager transition-based dependency parsing with Averaged perceptron and extended features
- Host: GitHub
- URL: https://github.com/touhi99/statistical_dependency_parsing
- Owner: touhi99
- Created: 2018-07-26T10:54:23.000Z (over 6 years ago)
- Default Branch: master
- Last Pushed: 2019-01-20T22:29:32.000Z (almost 6 years ago)
- Last Synced: 2023-11-26T21:28:18.876Z (12 months ago)
- Topics: averaged-perceptron, conll-2003, dependency-parsing, parse-trees, parsing, perceptron, transition
- Language: Python
- Homepage:
- Size: 5.86 KB
- Stars: 3
- Watchers: 2
- Forks: 0
- Open Issues: 0
-
Metadata Files:
- Readme: README.md
Awesome Lists containing this project
README
## Transition-based (Arc Eager) Dependency Parsing
### Author: Touhidul Alam
#### Usage:
```
python3 parser.py [train/dev/test] [en/de] file_path_location
```Example:
```
python3 parser.py train en /mount/studenten/dependency-parsing/data/english/train/wsj_train.only-projective.conll06
```### Dependencies:
* Numpy
* pickle### Components
* parser.py
* State - class initialize initial stack, buffer, arcs, left-dependency and right-dependency
* Instance - class saves the transition label with its corresponding features
* Parser - class does oracle parsing during training and predicting and making a parser* reader.py
* Corpus - class to read sentences from dataset and build them into sentence
* Sentence - holds the datasets content* model.py
* FeatureMapper - takes the unique feature and add it to the feature dictionary. Feature contains baseline feature, and extended features from Nivre (2011)
* Model - saves the FeatureMapper and weights, implemented with Averaged perceptron### Output File
```
pred_en.conll06 - English
pred_de.conll06 - German
```