关键词组鉴别的新方法——双向型高搭配力度ngram链

候选关键词组鉴别是关键词组自动化提取的第一步，在这个过程中采用一种新的数据结构：双向型高搭配力度ngram链。这种链式结构以搭配力度高者优先为延伸（生长）原则，可以向左右两个方向延伸（生长），经过对得到的ngrams消除冗余，可以得到一些结构比较规范的ngrams，可以作为文章的候选关键词组。

The new method of discriminating the key phrases-two-way high collocation ngram chain

The discrimination of the candidate key phrases is the first step of automatic retrieval of the key phrases, during which a new kind of new data structure is adopted, that is, two - way high collocation ngram chain. The chain structure follows the extension （growth） principle with priority to the high collocation and may extend in the direction of right and left. After the redundancy of the obtained ngrams is eliminated, the ngrams with comparatively normal structure may be obtained, which serve as the candidate key phrases.