Two Novel Decomposition Approaches for Knowledge Acquisition Model

摘要

Knowledge acquisition, one of essential issues for data mining, has always been a hot topic due to the explosive growth of information. However, when handling large-scale data, many current knowledge acquisition algorithms based on rough set theory are inefficient. In this paper, novel decomposition approaches for knowledge acquisition are put forward. The principal of decomposition is to split a complex problem in several problems. Those problems are composed of a master-problem and several sub-problems which are simpler, more manageable and more solvable by using existing induction methods, then joining them together in order to solve the original problem. Compared with some traditional algorithms, the efficiency of the proposed approaches can be illustrated by experiments with standard datasets from UCI database.