Sentiment Analysis for Malay News Headlines
Authors: Wong Wei Yee and Rayner Alfred
Many studies have been done in sentiment analysis which is focused on English documents. There are various sentiment analysis studies for movie or product reviews, blogs, and twitter messages being appeared and published. However, there are few studies on sentiment analysis performed on Malay documents. It is important to study sentiment analysis as it helps to study and understand the general mind-state of communities at a particular time with regard to some issue on news headlines. The Malay news headlines are few words and are often written with creativity to attract the readers’ attention. Sentiment analysis is a natural language processing technique which helps to identify the emotions, opinion and sentiments in any material source. Most sentiment analysis researches are done with help of supervised machine learning techniques. Therefore, this research is to design and implement the proposed supervised machine learning approach which is used to justify and classify sentiments on Malay news headlines into two categories: positive headlines and negative headlines. This research also aims to investigate the effects of using different classifier and various proximity measurements on the performance of the proposed approach of sentiment analysis for Malay news headlines. Based on the results obtained, Naïve Bayes classifier was capable to obtain high accuracy compared to k-NN classifier. For proximity measurement in k-NN classifier, based on the results obtained, it has shown that the cosine similarity outperform the Euclidean distance in terms of classification accuracy as the Euclidean distance measures can be affected by the high dimensionality of the data.