Learning to Recognize Affective Polarity in Similes
34920Authors
nan
Abstract
A simile is a comparison between two essentially unlike things, such as “Jane swims like a dolphin”. Similes often ex- press a positive or negative sentiment to- ward something, but recognizing the po- larity of a simile can depend heavily on world knowledge. For example, “memory like an elephant” is positive, but “mem- ory like a sieve” is negative. Our research explores methods to recognize the polarity of similes on Twitter. We train classifiers using lexical, semantic, and sentiment fea- tures, and experiment with both manu- ally and automatically generated training data. Our approach yields good perfor- mance at identifying positive and negative similes, and substantially outperforms ex- isting sentiment resources. 1
Keywords
Full Content
Proceedings of the 2015 Conference on Empirical Methods in Natural Language Processing, pages 190–200, Lisbon, Portugal, 17-21 September 2015. c⃝2015 Association for Computational Linguistics. Learning to Recognize Affective Polarity in Similes Ashequl Qadir and Ellen Riloff School of Computing University of Utah Salt Lake City, UT 84112, USA {asheq,riloff}@cs.utah.edu Marilyn A. Walker Natural Language & Dialogue Systems Lab University of California Santa Cruz Santa Cruz, CA 95064, USA mawalker@ucsc.edu Abstract A simile is a comparison between two essentially unlike things, such as “Jane swims like a dolphin”. Similes often ex- press a positive or negative sentiment to- ward something, but recognizing the po- larity of a simile can depend heavily on world knowledge. For example, “memory like an elephant” is positive, but “mem- ory like a sieve” is negative. Our research explores methods to recognize the polarity of similes on Twitter. We train classifiers using lexical, semantic, and sentiment fea- tures, and experiment with both manu- ally and automatically generated training data. Our approach yields good perfor- mance at identifying positive and negative similes, and substantially outperforms ex- isting sentiment resources. 1 Introduction A simile is a form of figurative language that com- pares two essentially unlike things (Paul, 1970), such as “Jane swims like a dolphin”. Similes of- ten express a positive or negative view toward an entity, object, or experience (Li et al., 2012; Fish- elov, 2007). Sometimes, the sentiment of a simile is expressed explicitly, such as “Jane swims beau- tifully like a dolphin!”. But in many cases the sen- timent is implicit, evoked entirely from the com- parison itself. “Jane swims like a dolphin” is eas- ily understood to be a compliment toward Jane’s swimming ability because dolphins are known to be excellent swimmers. A simile consists of four key components: the topic or tenor (subject of the comparison), the ve- hicle (object of the comparison), the event (act or state), and a comparator (usually “as”, “like”, or “than”) (Niculae and Danescu-Niculescu-Mizil, 2014). A property (shared attribute) can be op- tionally included as well (e.g., “He is as red as a tomato”). Our research aims to identify the af- fective polarity of a simile as positive, negative, or neutral, based on its component phrases. Table 1 shows examples of similes and their po- larity. A simile can have neutral polarity if it offers an objective observation. Example (a) is a neutral simile because, although bananas have a distinc- tive smell, it is not generally considered to be a particularly good or bad scent. Example (b) illus- trates that using the subjective verb “stink” instead of “smell” indicates a negative polarity toward the scent of bananas. Example (c) shows that includ- ing a subjective adjective such as “rotten” suggests a negative sentiment. Example (d) has negative polarity because the vehicle term, “garbage”, car- ries a strong negative connotation. Simile Polarity (a) smells like bananas neutral (b) stinks like bananas negative (c) smells like rotten bananas negative (d) smells like garbage negative (e) memory like an elephant positive …
Content truncated to 500 words
Labeled Text
A simile is a comparison between two essentially unlike things, such as “Jane swims like a dolphin”. Similes often express a positive or negative sentiment toward something, but recognizing the polarity of a simile can depend heavily on world knowledge. For example, “memory like an elephant” is positive, but “memory like a sieve” is negative. Our research explores methods to recognize the polarity of similes on Twitter. We train classifiers using lexical, semantic, and sentiment features, and experiment with both manually and automatically generated training data. Our approach yields good performance at identifying positive and negative similes, and substantially outperforms existing sentiment resources. 1
Publication Details
- Paper ID:
- 34920
- Year:
- 2015
- Publication:
- TIPSTER
- Publication Type:
- conf
- DOI:
- nan