A Top- Down Chart Parser for Ge’ez Sentences
No Thumbnail Available
Date
2024-12
Authors
Journal Title
Journal ISSN
Volume Title
Publisher
Addis Ababa University
Abstract
Parsing is the process of breaking down a sentence into its part of speech or words like verb, noun, preposition, adjective and so on. Parsing plays an important role in enhancing the performance of numerous natural languages processing (NLP). Here this work is designed to parse geez language on the approach of a Top-Down chart Parser for Geez language sentences using the context free grammar rule (CFG).
We reviewed various parsing approaches for different languages, to achieve our objective. How ever, Geez language parsing remains challenging dueto the absence of annotated dataset. To address this gap, we collected a dataset for sentence parsing from the well-known book Mezumere Dawit.Given the lack of the lack of pre-existing labled data; we collaborated with a language expert (Amanuel), an instructor in the Geez Department at Aksum University, to ensure lingustic accuracy. The expert prepared the dataset in a format suitable fstablished grammatical rules ford sentence construction based on verb and noun phrase structures, and manually parsed the sentences.
This study presents a Top-Down parsing approach for Geez language sentences; addressing the challenges posed by the language’s unique morphological and synthetic structures. Using a dataset of 500 sourced from the book ’Mezumure Dawit” the parser correctly parsed 470 sentences, achiving a parsing accuracy of 94%. The results were validated through the manual parsing, where 490 sentences were parsed manually, with 470 sentences matching the parser’s output. The performance of the parser was evaluated using standard metrics, including precision, recall, F1 score, and Experimental results show the effectiveness of the proposed method in parsing Geez language sentences. Thrid study contributes a foundational step towards computational processing of Geez language, with potential applications in machine translation and historical analysis.
Description
Keywords
CFG (Context Free Grammar), NLP (Natural language Processing), Top-Down Parser.