Improved algorithm for keywords extraction from documents without corpus

In this paper, an algorithm for extracting keywords without corpus is described. We use the co-occurrence information of the words and the biases of distribution to extract the more important words based on the most frequently appearing words so called reference words. Firstly, the most frequently terms are chosen from the document. Then due to keywords have a non-linear relationship with the set of frequently terms, the bias between words in documents and reference terms is measured. At last we prove that the algorithm is effective.