Association Rules Research And Analysis Of Parallel Algorithms

With the rapid development of information technology, it becomes more difficult and urgent to mine useful information and knowledge automatically in more and more larger amounts of data to support the strategies. The technology of data mining emerges under this background. Association Rule Mining is an important branch of data mining, and becomes one of the widest applied data mining styles. With the development of distributed database and the increase of data amount, research for parallel algorithm becoming the focus.The paper discusses and compares the traditional serial algorithms, analyses their virtues and disadvantages, then explained the necessary of parallel algorithm. This paper also introduces the cluster structure and discusses the model and methods of parallel programming. After analyzing the characteristic of some typical Parallel association rule algorithms, point out that there are also some problems to be solved.Parallel algorithms rely on the structure of parallel machine strongly, so the algorithm with good performance perhaps can not work well at another machine that has different structure. At the cluster system environment, we should use data parallel method to avoid masses communications. The paper proposes to use data parallel strategy based on Master/Slave model and analyses the performance.