DptOIE method uses the Dependence Parser and Part of Speech Tagger models trained with Stanford CoreNLP.
This work is described in the paper "DptOIE: A Portuguese Open Information Extraction system based on Dependency Analysis" under submission at Computer Speech and Language Journal.
- DptOIE's source code: An Open Information Extraction for Portuguese language.
- Dependenct Parser and Part of Speech Tagger models trained with Stanford CoreNLP.
- Dataset CETEN200, WIKI200 and models
- Insert the file pt-dep-parser.gz in pt-models directory
- Import as Maven Project in Eclipse IDE
To run the DptOIE.jar
java -jar DptOIE.jar -sentencesIN **sentences_file_path**
To use the module that handles subordinate clause
-SC true
To use the module that handles coordinated conjunctions
-CC true
To use the module that handles appositive
-appositive 1
To apply transitivity
-appositive 2
DptOIE is independent of dependency parser, so it can receive annotated sentences with other dependency parsers, as long as they are in ConLL-U format with the same tagsets of the Google treebank Treebanks Universal V2,1.
To run DptOIE from a dependency tree in ConLL-U format
java -jar DptOIE.jar -sentencesIN 'sentences_file_path' -dependencyTreeIN 'dependency_Tree_conllu_format'
Please use the GitHub issue tracker
- Leandro Souza de Oliveira
- Daniela Barreiro Claro
-
GLAUBER, R. ; CLARO, D. B. ; OLIVEIRA, L. S. . Dependency Parser on Open Information Extraction for Portuguese Texts - DptOIE and DependentIE on IberLEF. In: Iberian Languages Evaluation Forum (IberLEF 2019), 2019, Bilbao, Spain. Proceedings of the Iberian Languages Evaluation Forum (IberLEF 2019) co-located with 35th Conference of the Spanish Society for Natural Language Processing (SEPLN 2019). Bilbao: CEUR, 2019. v. 2421. p. 442-448.
-
Oliveira L., Claro D. B., Souza M., DptOIE: A Portuguese Open Information Extraction based on Dependency Analysis. Artificial Intelligence Review (under review)