基于混合方法的维吾尔语词干提取方法研究

Novel approach for Uyghur stemmer using mixed method

针对维吾尔语形态变化,提出了利用规则和词典相结合的混合处理方法进行形态还原技术。利用从左到右地分析和Lovin算法实现了词干提取器。通过总结词法连接规则,提出了规则实现词干提取、用词典验证提取结果。经过对不同新闻内容的五次测试得出平均准确率达到了77.4%。

This paper proposed changes in morphology of Uygur language, mixed processing method using a combination of rules and dictionaries phase morphology reduction technology. And proposed rules stemming and used a dictionary method to verify the extraction results. It are performed tests on the different combination of features. Experimental results show achieves recall of 77.4%.