一种改进的词汇语义极性分析算法*

词汇极性分析在自然语言处理等多个领域发挥着重要的作用。针对现有汉语词汇极性分析对词汇的义原考虑的不够全面,使得极性分析存在不准确的问题,提出一种更有效的词汇语义倾向性计算方法。首先找准一组褒贬中性词库,进而提取出待测词与词库中基准词词汇的极性义原、词性、其他义原,计算相应的相似度;并分别给出合适的权重值。由此计算出待测词与基准词的相似度,最后判断待测词的语义极性。实验表明该算法可以进一步的提高语义极性倾向判别准确率。

An Improved Algorithm for Lexical Semantic Polarity Analysis

Glossary polar analysis plays an important role in a number of areas such as the natural language processing. For the analysis of the existing Chinese vocabulary words polar sememe considered incomplete, making polarity analysis inaccurate enough, then a higher propensity of a lexical semantic precision calculations is proposed. Firstly, the polarity sememe, part of speech,and the other sememes of the under test word and the standard word in praise-derogatory- neutral library are extracted, the corresponding similarity of them are calculated by using designed formulas. Then, the similarity of the under test word and the standard word can be got through by giving appropriate weight value to the corresponding similarity. At last, the semantic polarity of the under test word can be judgment. Experimental results show that the proposed method can further improve the accuracy of semantic polarity discrimination tendencies.