A frequent pattern parallel mining algorithm based on distributed sliding window

摘要

Cloud computing infrastructure based on Map/Reduce model provided a promising solution to solve Association Rule Mining of large data, but redundancy of candidate sets and low efficiency of frequent pattern mining were still important problems in distributed data mining. A parallel algorithm based on Hadoop cloud computing platform is proposed, which constructs and mines TPT-Tree in distributed nodes and processes intermediate candidate sets with hash structure to accelerate the speed of mining frequent item sets. Experimental results show that this algorithm improves the efficiency of frequent pattern data mining and is with a satisfied speedup.