藏语及物与不及物动词的自动识别研究

藏语动词分类很细,准确识别和判断藏文动词的词性是藏语自然语言处理中一个基础又关键的内容,及物与不及物动词的自动识别,可以在用计算机生成自然语言句子时能很好地阻止不合法句子的生成,也能从整体上减少句子生成量,从而提高效率.提出了一种在理想环境下,通过句子中的使格助词来自动判断动词及物性的方法,这种自动识别方法是一种理性主义的判断方法,所以不需要大规模语料的支持.因为是一种理想情况,所以自动识别效率高,准确率为100%.

Research on the automatic identification of Tibetan transitive and intransitive verbs

Tibetan verb has a very fine classification, so an accurate identification and judgment of Tibetan verb＇s parts of speech is a foundation in the Tibetan natural language processing and key task. An automatic identification of transitive and intransitive verbs can enable the computer to prevent the formation of illegal sentences while generating natural language sentences, and also can help reduce the number of the sentences to be generated as a whole, so as to improve efficiency. The paper proposes a method which, under an ideal environment, can automatically determine the verb in the sentence and the property by using case of utterances. This method of automatic identification is a judgment method of rationalism which does not need the support of large-scale corpus. Because it is an ideal situation, the automatic identification is of high efficiency with an accuracy of 100%.