一种多层次分布式网络数据挖掘方法的改进研究

针对传统数据挖掘方法时间开销大的问题,提出一种新的多层次分布式网络数据挖掘改进方法,给出多层次分布式网络结构。介绍了常用的随机扰动数据挖掘方法,通过概率歪曲技术完成对多层次分布式网络中原始数据集的扰乱处理,对项集的实际支持度进行重构,经概率转换获取数据挖掘结果。随机扰动方法具有时间效率低的弊端,在时间复杂度要求较低的情况下,通过XMASK方法对随机扰动方法进行改进;在时间复杂度要求较高的情况下,给出相应的改进过程。对提出的多层次分布式网络数据挖掘方法进行实验测试,结果表明,该方法准确性高、挖掘时间短、效果优。

Research on an improved method of multi level distributed network data mining

Aiming at the problem of high cost of traditional data mining methods, this paper proposes a new method to improve the multi-level distributed network data mining. The random disturbance data mining method commonly used, through the probability distorted technology to disrupt the original data processing multi-level distributed network set, to reconstruct the actual item set support, the probability of conversion and acquisition of data mining results. Random perturbation method has the disadvantages of low time efficiency, the time complexity is low, the random perturbation method improved by XMASK method; the time complexity is high, the corresponding improvement process. The experimental results show that the proposed method has high accuracy, short mining time and good effect.