Classifying COVID-19 News on Sina Weibo

Abstract

This thesis addresses the classification of Sina Weibo news related to the COVID-19 pandemic by using sentiment analysis. The design is a comparison study, involving four different systems. The systems were chosen after extensive reading on different approaches to sentiment analysis in Mandarin Chinese. These systems are neural network, k-medoids, HMM, and SVM. The core work of this thesis was in fine-tuning these systems to work with a small dataset of less than 3,000 examples. The final results from numerous experiments showed that the SVM and HMM systems achieved the highest results, followed by the neural network and k-medoids systems. Findings of this study showed that keyword frequency within a news category is not necessarily sufficient to ensure correct classification. Also, posts with names and satire remain challenging to classify and could be investigated further.