The Semantic Knowledge Acquisition Of Chinese Unknown Words

In natural language processing (NLP), it is a basic work to query the correlative information of a word from a lexicon, and the semantic acquisition of a word is a necessary process. Under the condition of a great of information, the dynamic and the real time, unknown words sometimes offer very important information. Since we cannot acquire the semantic knowledge directly from a lexicon and we must guess the meaning of unknown words relying on the morphological structure and the contexts, it becomes a very significant and valuable field of research.In order to acquire the semantic knowledge of unknown words, we do some research work as follows:1.The definition of Chinese unknown words: To define Chinese unknown words is the foundation of semantic acquisition. Through analyzing other correlative work, we choose "2+1" compound new words as the research object of this paper.2.Accrording to the morphological structure of Chinese unknown words, we promote a method of semantic classification based on Cilin. By means of computing the semantic similarity of unknown words and Cilin words which has the same morphological structure as the unknown word, we put the unknown word into the semantic category of Cilin word which is the most similar to it.3. This paper promotes a kind of measure used to compute the semantic similarity of words — considering the contextual information of the words ,describing the semantic knowledge of words by using contextual word co-occurrence vector (CWCV) and on the basis of the relation between the similarity and the association of the words, computing the semantic similarity of words by using Min/Max measure.4.According to the deficiency of artificial evaluation, we promote a evaluating method which does not need people to intervene. We take on the Cilin words as the unknown words whose semantic category is unknown,...