ESURF: Simple and Effective EDU Segmentation
Computation and Language
2025-01-15 v1 Machine Learning
Abstract
Segmenting text into Elemental Discourse Units (EDUs) is a fundamental task in discourse parsing. We present a new simple method for identifying EDU boundaries, and hence segmenting them, based on lexical and character n-gram features, using random forest classification. We show that the method, despite its simplicity, outperforms other methods both for segmentation and within a state of the art discourse parser. This indicates the importance of such features for identifying basic discourse elements, pointing towards potentially more training-efficient methods for discourse analysis.
Cite
@article{arxiv.2501.07723,
title = {ESURF: Simple and Effective EDU Segmentation},
author = {Mohammadreza Sediqin and Shlomo Engelson Argamon},
journal= {arXiv preprint arXiv:2501.07723},
year = {2025}
}