Abstract:

Keywords extraction is very important for the text similarity computing.By the traditional keywords extraction algorithm,the unknown words in the text are ignored and word semantics concept is lacking.In this paper,an extraction algorithm based on unknown words recognition and semantic is proposed for technology project applications.The algorithm integrates Lucene and statistics method for word segmentation and unknown words recognition,and a part of keywords are achieved.Then,the other keywords of applications are extracted by constructing the word semantic similarity network which based on Social Network and calculating the associated degrees of words.Experimental results show that compared with the traditional keyword extraction algorithm,the proposed algorithm can extract more accurate keywords,and it provides a better result of checking on the similarity of technology projects.

Key words: keywords extraction;unknown words;social network theory;semantic similarity network