Comparative Study of Vietnamese Part-of-Speech Tagging Tools

• Main Authors: Quach, L.-D., Do Thanh, D., Tran, D.C., Hassan, M.F.
• Format: Conference or Workshop Item
• Published: Institute of Electrical and Electronics Engineers Inc. 2020
• Online Access: https://www.scopus.com/inward/record.uri?eid=2-s2.0-85096933078&doi=10.1109%2fICSGRC49013.2020.9232564&partnerID=40&md5=7b4f98653c96712391c87ba08ced7cc7; http://eprints.utp.edu.my/30115/
• Institution: Universiti Teknologi Petronas

Vietnamese part-of-speech tagging is one of the most fundamental practices in Vietnamese language processing. Unfortunately, no attempt has been made to empirically compare different Vietnamese part-of-speech tagging software. Therefore, in this paper, the authors experiment upon several Vietnamese part-of-speech tagging software such as VnTagger, RDRPOSTagger (Java Version), JvnTextPro, VNCoreNLP in terms of accuracy, consistency and computational time. In addition, the brief descriptions of the models are discussed in detail. The results help researchers comprehend the models' strengths and weaknesses. The tools are tested on 4 different data sets of number of sentences and different word types such as date, number, special characters, connected characters, double words, compound words, proper names, etc� The results show that the accuracy of the JvnTextPro tool is high and stable with an accuracy of 80.08 to 97.84, and the RDPRPOSTagger tool has faster processing time and relatively good accuracy from 88.41 to 96.84. © 2020 IEEE.