Multi-resolution Keyword Mining Algorithm based on Frequent Pattern Technique
Xiaodong Ji · 2010
Finding out keywords from massive text document is a hard problem because an amount of synonyms are frequently used in a text.This paper proposes a keyword mining algorithm based on analyzing the "frequent patterns" of the words in a text.We use a new method referring the cluster process to determine the words' relationship by their distances.We firstly extend the FP-tree to a new structure named W3P-Tree which is used to maintain the text information instead of database transactions, then present our mining algorithm following with the growth of W3P-tree.