Salvato in:
Dettagli Bibliografici
Autori principali: Salehi, Ali, Jacobs, Cassandra L.
Natura: Preprint
Pubblicazione: 2025
Soggetti:
Accesso online:https://arxiv.org/abs/2511.14696
Tags: Aggiungi Tag
Nessun Tag, puoi essere il primo ad aggiungerne!!
_version_ 1866912716771295232
author Salehi, Ali
Jacobs, Cassandra L.
author_facet Salehi, Ali
Jacobs, Cassandra L.
contents We investigate tokenization strategies for Kurdish word embeddings by comparing word-level, morpheme-based, and BPE approaches on morphological similarity preservation tasks. We develop a BiLSTM-CRF morphological segmenter using bootstrapped training from minimal manual annotation and evaluate Word2Vec embeddings across comprehensive metrics including similarity preservation, clustering quality, and semantic organization. Our analysis reveals critical evaluation biases in tokenization comparison. While BPE initially appears superior in morphological similarity, it evaluates only 28.6\% of test cases compared to 68.7\% for morpheme model, creating artificial performance inflation. When assessed comprehensively, morpheme-based tokenization demonstrates superior embedding space organization, better semantic neighborhood structure, and more balanced coverage across morphological complexity levels. These findings highlight the importance of coverage-aware evaluation in low-resource language processing and offers different tokenization methods for low-resourced language processing.
format Preprint
id arxiv_https___arxiv_org_abs_2511_14696
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Subword Tokenization Strategies for Kurdish Word Embeddings
Salehi, Ali
Jacobs, Cassandra L.
Computation and Language
We investigate tokenization strategies for Kurdish word embeddings by comparing word-level, morpheme-based, and BPE approaches on morphological similarity preservation tasks. We develop a BiLSTM-CRF morphological segmenter using bootstrapped training from minimal manual annotation and evaluate Word2Vec embeddings across comprehensive metrics including similarity preservation, clustering quality, and semantic organization. Our analysis reveals critical evaluation biases in tokenization comparison. While BPE initially appears superior in morphological similarity, it evaluates only 28.6\% of test cases compared to 68.7\% for morpheme model, creating artificial performance inflation. When assessed comprehensively, morpheme-based tokenization demonstrates superior embedding space organization, better semantic neighborhood structure, and more balanced coverage across morphological complexity levels. These findings highlight the importance of coverage-aware evaluation in low-resource language processing and offers different tokenization methods for low-resourced language processing.
title Subword Tokenization Strategies for Kurdish Word Embeddings
topic Computation and Language
url https://arxiv.org/abs/2511.14696