Aspect-Based Sentiment Analysis for Posts on Friday Prayer During MCO in Malaysia

Abstract

Analysis of sentiment (or opinion mining) is a technique used to determine whether a polarity of data has become positive, negative, or neutral. It studies the opinions, feelings, emotions, and stances of people using an algorithmic process that understands the opinions of a particular topic based on the methodology of Natural Language Processing (NLP). It has gained popularity in recent years and it has played a vital role in a variety of fields, such as online product reviews and social media analysis (Twitter, Facebook, etc.). This paper presents the findings of a research conducted to investigate people’s sentiment toward a government decision that temporarily suspending Friday prayers in all the mosques, as a response to the pandemic of COVID-19 in the country, due to The Malaysia Movement Control Order (MCO) 1.0 as a precautionary measure. A collection of tweets were crawled based on the #solatjumaat hashtag, then it was grouped into one corpus as a new dataset for further text preprocessing and sentiment analysis process. It applies a Python language with an adaption of Malaya, a Natural-Language-Toolkit library created especially for text in Malay Language verse for the treatment techniques. A visualization of the outcome will illustrate the finding of people's feelings for this study.