An application of Malay short-form word conversion using Levenshtein distance / Azilawati Azizan, NurAine Saidin, Nurkhairizan Khairudin & Rohana Ismail

• Main Authors: Azilawati Azizan, Azilawati Azizan, NurAine Saidin, NurAine Saidin, Nurkhairizan Khairudin, Nurkhairizan Khairudin, Rohana Ismail, Rohana Ismail
• Format: Article
• Language: English
• Published: UiTM Press 2020
• Subjects: Malaysia; Algorithms
• Online Access: https://ir.uitm.edu.my/id/eprint/38191/2/38191.pdf; https://ir.uitm.edu.my/id/eprint/38191/; https://mijuitm.com.my/view-articles/
• Institution: Universiti Teknologi Mara

Formerly, short-form word was widely used in the field of journalism. However, nowadays, short-form word has been widely used by many people, especially in online communication. These short-form words trigger problems in the field of data mining, especially those involving online text processing. It leads to inaccurate result of text mining activities. On the other hand, only few works have investigated on Malay short-form word identification and conversion. Therefore, this work aims to develop an application that can identify and convert Malay short-form words into its’ full word. In order to develop this application, the short-form rules need to be carefully examined. The formal rules from Dewan Bahasa & Pustaka (DBP) are used as the primary reference for generating the short form word identification algorithm. While for the conversion algorithm, Levenshtein Distance (LD) is used to measure the similarity. The rule-based technique is also used as a complement to LD technique. As a result, 70.27% of the Malay short-form words have been correctly converted into their full words. The conversion rate is quite promising, and this work can be further strengthened by incorporating more rules into the algorithm.