Abstract:
Extraction of Keywords using Frequent Itemsets (AEKFI) is a new technique for keyword extraction which integrates adjacency of location of words within the document to automatically select the most discriminative words without using a corpus. This paper introduces a novel Binomial Heap Approach based AEKFI for document summarization. Binomial heap does keyword extraction using binomial minimum heap operations. AEKFI provides flexibility to select either the set of keywords from a given document or user specified number of keywords. AEKFI does not impose any restriction on the length of keywords being extracted. Demonstration of Binomial Heap Extractor has been made and has been found efficient in reducing the time complexity O (n2) of existing approaches to O (n log n). Experimental results prove the advantage of Binomial Minimum Heap based AEKFI over other keyword extraction tools.
