一种Apriori算法的改进

An Improved Apriori Algorithm

Apriori算法在处理关联规则分析时,当数据立方体数据稠密时,实现迭代性质将需要非常复杂的数据结构。针对上述问题,本文提出了一种改进的Apriori-ni算法,该算法没有用迭代性质来剪枝,即不基于迭代属性的算法。对Apriori算法和Apriori-ni算法进行了分析和比较,实验结果表明,当项目集很多时,Apriori-ni算法能节约计算开销,从而提高算法的效率。

When the Apriori algorithm deals with the relevant rule analysis and the data-cube＇s data is dense, it needs very complex data structures to realize the iteration characteristics. According to the aboved problems, we propose an improved Apriori-ni algorithm, which does not make use of the iteration characteristics to prune and does not bdong to the algorithm of the iteration characteristics. Through the comparision between the Apriori algorithm and the Apriori-ni algorithm, the results indicate that the later one can save the caculation cost , and thus improve the algorithm effieieney.