The Utility Frequent Pattern Mining Based On Slide Window In Data Stream

With the high development of hardware, people’s ability in collecting data is aswell in fast improvement. More and more fields can no more meet the demands ofprocessing and analyzing the static data, hence, the searching hotspot for datamining is transferring into data streams. Differ from the static data, data stream hasthe fast, multi-changing, living and magnanimity characteristics which bring thenew difficulty and challenge for data mining work. In recent research of datastream mining, each item in data stream is of equal importance. However, inpractice, each item has a different significance-that is utility which this method iscalled frequent pattern mining. However, some theorem and methods are no more inadaption, like downward close space which bring a huge difficulty in mining tasks.This paper combines frequent mining item sets with utility and proposes anefficient algorithm for utility frequent pattern mining (UFPM). It combines bitmapwith tree structure that can store and update the pattern of data stream quickly andcompletely by scanning only once. The algorithm generated by lexicographic order,proposes a novel tree U-tree and makes convenience for pattern updating and userreading. With a pattern growth approach in mining, the algorithm can effectivelyavoid the problem of a mass candidacy generation by level-wise searching. Byworking through several experiments to testing and contrasting in existingalgorithms, we prove that our algorithm is in high efficiency and good scalabilityoutperforms the existing analogous algorithm.