Generating a Malay sentiment lexicon based on wordnet

Abstract

Sentiment lexicon is a list of vocabularies that consists of positive and negative words. In opinion mining, sentiment lexicon is one of the important source in text polarity classification task in sentiment analysis model. Studies in Malay sentiment analysis is increasing since the volume of sentiment data is growing on social media. Therefore, requirement in Malay sentiment lexicon is high. However, Malay sentiment lexicon development is a difficult task due to the scarcity of Malay language resource. Thus, various approaches and techniques are used to generate sentiment lexicon. The objective of this paper is to develop Malay sentiment lexicon generation algorithm based on WordNet. In this study, the method is to map the WordNet Bahasa with English WordNet to get the offset value of a seed set of sentiment words. The seed set is used to generate the synonym and antonym semantic relation in English WordNet. The highest result achives 86.58% agreement with human annotators and 91.31% F1-measure in word polarity classification. The result shows the effectiveness of the proposed algorithm to generate Malay sentiment lexicon based on WordNet.