基于二叉频繁模式树的医学图像关联规则挖掘

Associative Regular Mining Based on Frequent Pattern Bintree for Medical Images

医学图像的数据量是相当巨大的,挖掘医学图像中数据的关联关系就需要一种适合挖掘海量数据的挖掘算法。针对基于频繁模式树FP_TREE的关联规则算法在挖掘海量数据时占用大量内存的缺点,提出了一种基于二叉频繁模式树(FP_BTREE)的关联规则算法。该算法采用二叉树存储数据的技术来映射数据库中数据,以减少对数据库的访问次数。而且根据内存具体情况可以先求出先建立的二叉频繁模式树的频繁模式。解决了占用大量内存的缺点,适合挖掘医学图像海量数据集。此算法也为多棵二叉频繁模式树的并行计算打下基础。最后应用此算法提取医学图像数据集中隐含的关联信息。

The amount of data of medical image is quite huge.To mine the connected relationship among medical image,we need a algorithm which is suitable to huge data.Taking too much memory is the disadvantage of frequent pattern tree FP_tree,when its association regular method is used to mine maganimity data.So we present a association rule mining algorithm based on FP_Btree,which uses the technology of bintree data to reflect the data of data sets.This method covers a little memory,which is suitable to extract medical image data sets.This algorithm lays the basis on the parallel calculation of the FP-Btree.Finally,with the method,it can also extract the hidden associated information from the medical image data