Distributed pruning optimization oriented FP-Growth method based on PSO algorithm

摘要

Due to such problems as low-efficiency and low-precision using the traditional FP-Growth algorithm with huge amount of data, this paper raises MSPF algorithm based Particle swarm optimization. This method used anti-monotony property pruning approach of condition tree to reduce the searching space and increase mining efficiency. At the same time, using the distributed calculating platform Hadoop and distributed computing framework MapReduce parallelize the MSPF algorithm which named PMSPF algorithm. The experiment is showing is that improved distributed algorithm has a better performance than the traditional FP-Growth algorithm and PFP-Growth algorithm.