AN ALGORITHM FOR DATA MINING ON LARGE INCEMENT DATABASE BASED ON THE PARTITION DATA
Nguyễn Hữu Trọng
Abstract
With all algorithms for datamining on fixed database were announced, whenever the dataincrements, the problem has to be solved from the beginning again; whenever a frequent itemsetis sought with another support, the seeking needs to be start again. Some algorithms fordatamining on'incrementing database were announced. This paper presents an algorithm forsolving the dala mining problems on large incrementing database, based on the vertical dataformat. Whenever the data increments or solving with another support, the algorithm onlyprocesses on new transactions and connected to the previous result without going back to thestarting point.