Morpheme-based Derivation of Bipolar Semantic Orientation of Chinese Words

Morpheme-based Derivation of Bipolar Semantic Orientation of Chinese Words Raymond W.M. Yuen, Terence Y.W. Chan, Tom B.Y. Lai, O.Y. Kwong, Benjamin K.Y. T'sou Language Information Sciences Research Centre, the City University of Hong Kong 83 Tat Chee Avenue, Hong Kong { wmyuen, dcywchan, cttomlai, rlolivia, rlbtsou}@cityu.edu.hk Abstract The evaluative character of a word is called its semantic orientation (SO). A positive SO indicates desirability (e.g. Good, Honest) and a negative SO indicates undesirability (e.g., Bad, Ugly). This paper presents a method, based on Turney (2003), for inferring the SO of a word from its statistical association with strongly-polarized words and morphemes in Chinese. It is noted that morphemes are much less numerous than words, and that also a small number of fundamental morphemes may be used in the modified system to great advantage. The algorithm was tested on 1,249 words (604 positive and 645 negative) in a corpus of 34 million words, and was run with 20 and 40 polarized words respectively, giving a high precision (79.96% to 81.05%), but a low recall (45.56% to 59.57%). The algorithm was then run with 20 polarized morphemes, or single characters, in the same corpus, giving a high precision of 80.23% and a high recall of 85.03%. We concluded that morphemes in Chinese, as in any language, constitute a dis- tinct sub-lexical unit which, though small in number, has greater linguistic significance than words, as seen by the significant en- hancement of results with a much smaller corpus than that required by Turney. 1. Introduction The semantic orientation (SO) of a word indicates the direction in which the word deviates from the norm for its semantic group or lexical field (Lehrer, 1974). Words that encode a desirable state (e.g., beautiful) have a positive SO, while words that represent undesirable states (e.g. absurd) have a negative SO (Hatzivassiloglou and Wiebe, 2000). Hatzivassiloglou and Mckeown (1997) used the words ‘and’, ‘or’, and ‘but’ as linguistic cues to extract adjective pairs. Turney (2003) assessed the SO of words using their occurrences near strongly- polarized words like ‘excellent’ and ‘poor’ with accuracy from 61% to 82%, subject to corpus size. Turney’s algorithm requires a colossal corpus (hundred billion words) indexed by the AltaVista search engine in his experiment. Undoubtedly, internet texts have formed a very large and easily- accessible corpus. However, Chinese texts in internet are not segmented so it is not cost- effective to use them. This paper presents a general strategy for inferring SO for Chinese words from their association with some strongly-polarized morphemes. The modified system of using morphemes was proved to be more effective than strongly-polarized words in a much smaller corpus. Related work and potential applications of SO are discussed in section 2. Section 3 illustrates one of the methods of Turney’s model for inferring SO, namely, Pointwise Mutual Information (PMI), based on the hypothesis that the SO of a word tends to correspond to the SO of its neighbours. The experiment with polarized words is presented in section 4. The test …

Content truncated to 500 words