Conducting Sentiment Analysis. Lei Lei and Dinlin Liu

All human languages, whether spoken or written, transmits sentiment, and emotion, which is a major field of linguistic research. Researchers used to examine this issue mainly in terms of semantic prosody, which describes a word that commonly co-occurs with other words that correspond to a specific positive/neutral/negative semantic set. Such studies use context to infer the sentiment or emotion of single words, but the challenge now is determining the sentiment orientation of words, phrases, and even entire texts in documents with millions of words or more. Researchers used to manually tag negative and positive words in texts in early studies, which was time-consuming and labor-intensive. Later, as large-scale corpora and corpus linguistics advanced, data to anyone looking for solid empirical evidence to test, illustrate, and demonstrate linguistic hypotheses and theories. Sentiment analysis is being increasingly integrated with linguistics, particularly corpus-based linguistics, and is gaining popularity in corpus-based language teaching, discourse analysis, translation research, and other domains as a form of natural (NLP).