Sentiment Polarity Classification for Khmer

Sentiment polarity classification is a natural language processing (NLP) technique that determines whether a given piece of text has a positive, negative, or neutral sentiment. This article investigates the application of a well-known shallow neural network classifier, FastText, to the Khmer language, which is a low-resource language. To our knowledge, there isn’t a freely available polarity corpus for Khmer. Thus, one of the primary objectives of this research is to develop a human-annotated benchmark corpus tailored for sentiment analysis in this language. Our findings indicate that the FastText model achieves a precision and recall of 0.740 using our modest-sized Khmer polarity corpus, which consists of 9,000 training sentences and 1,000 testing sentences. Among traditional machine learning methods, the support vector machine (SVM) outperforms the rest.