An incremental attribute reduction algorithm for distributed decision information system

How to effectively carry out attribute reduction is an important task when some data vary dynamically in the decision system, many incremental attribute reduction methods have been proposed. Whereas there is less work on incremental attribute reduction method for distributed decision information system. In order to address this problem, a matrix-based incremental attribute reduction algorithm is proposed for distributed data. In this paper, we first introduce some definitions and conceptions of distributed decision information system. Then, the matrix-based incremental mechanisms for knowledge granularity are discussed and the corresponding matrix-based incremental attribute reduction algorithm for updating reduct is developed. Finally, some data sets is downloaded from the UCI machine learning libraries and some numerical experiments demonstrate that the incremental attribute reduction algorithm is effective and efficient compared to non-incremental attribute reduction algorithm.