Context-Enhanced Citation Sentiment Detection
Awais Athar, Simone Teufel · 2012
Sentiment analysis of citations in scientific papers and articles is a new and interesting problem which can open up many exciting new applications in bibliographic search and bibliometrics. Current work on citation sentiment detection focuses on only the citation sentence. In this paper, we address the problem of context-enhanced citation sentiment detection. We present a new citation sentiment corpus which has been annotated to take the dominant sentiment in the entire citation context into account. We believe that this gold standard is closer to the truth than annotation that looks only at the citation sentence itself. We then explore the effect of context windows of different lengths on the performance of a stateof-the-art citation sentiment detection system when using this context-enhanced gold standard definition. However, there is a problem with the expression of sentiment in scientific text. Conventionally, the writing style in scientific writing is meant to be objective. Any personal bias by authors has to be hedged (Hyland, 1995). Negative sentiment is politically particularly dangerous (Ziman, 1968), and some authors have documented the strategy of prefacing the intended criticism by slightly disingenuous praise (MacRoberts and MacRoberts, 1984). This makes the problem of identifying such opinions particularly challenging. This non-local expression of sentiment has been observed in other genres as well (Wilson et al., 2009; Polanyi and Zaenen, 2006). 1