Please use this identifier to cite or link to this item: https://hdl.handle.net/10356/76125
Title: Singlish concept-level sentiment analysis
Authors: Piyawan, Andrew Yew Ming Preedee
Keywords: DRNTU::Engineering::Computer science and engineering
Issue Date: 2018
Abstract: With a steady increment of people using social media platforms such as Facebook and Twitter as a form of communication over the recent years, a substantial amount of data is being generated daily. Through the data collected from various social media platforms daily, companies can understand customers’ opinions on their new and existing products, and services provided by them. However, to analyse a substantial amount of data that is increasing exponentially every minute could be a challenging problem. Singapore is known for being a multiracial and multicultural country; it is diverse in its languages. The four official languages of Singapore are English, Mandarin, Malay and Tamil respectively. While English is not only used in school and work but also acts as a bridge of communication and interaction between the various ethnic groups in a multicultural society. The use of English is generally categorized into two unique forms – Singapore Standard English and Singapore Colloquial English (also known as Singlish). In general, the former is mainly used in formal contexts such as presentation whereas the latter is widely popular in informal communication such as social media or daily conversation. In order to have a better understanding of the data that is readily available, there is a need to extend and grow towards building colloquial languages as multilingual sentiment analysis have been gaining attention but currently, it is limited to standardized languages such as English and other languages such as Chinese and Spanish. In this study, we have come up with Facebook messenger Chatbot and with the inputs from native Singlish speakers, we are able to extend Singlish knowledge base by gathering even more Singlish concepts that are important and useful for this study. In addition, a preliminary Singlish polarity analyser that returns the polarity (positive or negative) of Facebook comments, which achieved an accuracy of 74%.
URI: http://hdl.handle.net/10356/76125
Rights: Nanyang Technological University
Fulltext Permission: restricted
Fulltext Availability: With Fulltext
Appears in Collections:SCSE Student Reports (FYP/IA/PA/PI)

Files in This Item:
File Description SizeFormat 
Andrew FYP Report.pdf
  Restricted Access
927.9 kBAdobe PDFView/Open

Page view(s) 1

167
checked on Oct 27, 2020

Download(s) 1

55
checked on Oct 27, 2020

Google ScholarTM

Check

Items in DR-NTU are protected by copyright, with all rights reserved, unless otherwise indicated.