增量式目标信息系统的分布约简算法

摘要

Knowledge reduction is one of the important research contents in rough set theory.Static distribution reduction algorithm is not adapted to dealing with the information system in rapid growth, and it is very time-consuming.Through studying the impact of newly incre-mented object on the division of information systems, we discussed the theoretical relationship between the objects and the distribution reduc-tion by taking division as the bridge.Based on existing distribution reduction algorithm, we proposed the new distribution reduction algorithm for incremental object information system, made it to be able to combine the existing knowledge with newly incremented data, and to rapidly obtain the distribution reduction.The data sets downloaded from UCI databases was used to test the algorithm, it verified that the incremental algorithm is effective and feasible.%知识约简是粗糙集理论中的重要研究内容之一。由于静态分布约简算法不适应处理快速增长的信息系统，时间开销过大。通过研究新增对象对于信息系统中划分的影响，以划分为桥梁，讨论对象与分布约简的理论关系。基于已有的分布约简算法，提出增量式目标信息系统的分布约简算法，使其能够将原有知识和新增数据相结合，快速获取分布约简。使用UCI数据库中的数据集对算法进行测试，验证了该增量式算法的有效性和可行性。