A Review of Sentiment Analysis Research in Chinese Language
Haiyun Peng,Erik Cambria,Amir Hussain
DOI: https://doi.org/10.1007/s12559-017-9470-8
IF: 4.89
2017-05-08
Cognitive Computation
Abstract:Research on sentiment analysis in English language has undergone major developments in recent years. Chinese sentiment analysis research, however, has not evolved significantly despite the exponential growth of Chinese e-business and e-markets. This review paper aims to study past, present, and future of Chinese sentiment analysis from both monolingual and multilingual perspectives. The constructions of sentiment corpora and lexica are first introduced and summarized. Following, a survey of monolingual sentiment classification in Chinese via three different classification frameworks is conducted. Finally, sentiment classification based on the multilingual approach is introduced. After an overview of the literature, we propose that a more human-like (cognitive) representation of Chinese concepts and their inter-connections could overcome the scarceness of available resources and, hence, improve the state of the art. With the increasing expansion of Chinese language on the Web, sentiment analysis in Chinese is becoming an increasingly important research field. Concept-level sentiment analysis, in particular, is an exciting yet challenging direction for such research field which holds great promise for the future.
computer science, artificial intelligence,neurosciences