A High Performance Frequent Itemset Mining Algorithm Using Confidence Frequent Pattern Tree

Various processing methods for association data mining are presently being looked into. Most of them focus on data structure and computation improvement. The data structures usually have a high degree of data compression ratio and can express the original information from the database with integrity. There is also no need to obtain information from the database again. However, not many studies concentrate on using known frequent item sets to increase system performance. In order to avoid repeating the calculation of known frequent items to speed up the data mining process, a new tree structure to store all known frequent item sets and a header table to create a frequent item linking list are proposed. The experimental results showed that the proposed procedure performs better compared with existing data mining procedures.