OOV_Handling_for_Dependency_Parsing

---

This is the python implementaion of our work here: "SOON"

Requirements

Python 3

requirements.txt:
tqdm==4.30.0
scikit-learn==1.0
nltk==3.4.5
torch==1.6.0
pytorch-pretrained-bert==0.4.0
seaborn==0.9.0

Datasets

---

AR ArPoT: here

SEV Method for Replacing OOV Poetry Vocabulary (Unseen by Pre-Trained BERT)

• S: Synonyms
• E: Embeddings
• V: Variants

The replacements with known words for each level are available under the folder: OOV_Replacements

Parsing Experiments

The codes for parsing experiments are available under the folder: Python_Codes

Parsing as Sequence labeling

In our work, we have adopted the approach outlined by Vilares et al., utilizing their code as a foundation. Additionally, it is pertinent to acknowledge the contributions of Strzyz et al. for their influential work in the field. Our methodology also aligns with the findings presented in our research article available on MDPI, which further substantiates our approach.

Citation

---

Soon

# Contact information
---
For help or issues send an email to sharefah@ksu.edu.sa.