Analysis on Li Qingzhao’s Poems: Discussion on the Limitations of Applying Sentiment Analysis on Chinese Literatures of the Song Dynasty

Conference: The Asian Undergraduate Research Symposium (AURS2021)
Title: Analysis on Li Qingzhao’s Poems: Discussion on the Limitations of Applying Sentiment Analysis on Chinese Literatures of the Song Dynasty
Stream: Science, Technology, Engineering & Math
Presentation Type: Virtual Poster Presentation
Authors:
Hoi Ting Law, The University of Hong Kong, Hong Kong

Abstract:

Natural language processing (NLP), concerning the interactions between computers and human languages, has been widely used in processing and analyzing large amounts of data in human languages. NLP not only enables computers to understand the contents of human languages in different media, for example documents and audio, its subfields have been universally employed to handle lexical tasks. Sentiment analysis, which has been used to quantify and study affective states and subjective information, is an example. Although a rising number of packages and algorithms for sentiment analysis have been introduced to further enhance the accuracy and usefulness of the analysis, they are not that developed for analyzing Chinese characters compared to that of English. SnowNLP, one of the few packages for sentiment analysis in Chinese, despite being acclaimed as having satisfying accuracy in the analysis, its efficiency in analyzing ancient Chinese characters is still questionable. This paper aims to evaluate the efficiency and accuracy of sentiment analysis on classical in Chinese literature of the Song Dynasty by using SnowNLP to analyze all poems (Song Ci) of Li Qingzhao, a famous female poet in the Song Dynasty. In this paper, an experimental analysis of all poems of Li by SnowNLP are discussed, followed by the evaluation based on the analysis results. Finally, suggestions on ameliorating the outcomes of sentiment analysis are provided.



Virtual Presentation



Virtual Poster Presentation

virtual-poster-62029

Conference Comments & Feedback

Place a comment using your LinkedIn profile

Comments

Share on activity feed

Powered by WP LinkPress


Share this Presentation