论文部分内容阅读
粗糙集理论被广泛应用于属性约简,算法复杂性是制约约简应用于大样本知识发现的主要问题,尤其是邻域模型下的约简问题.本文分析邻域粗糙集模型的数学性质,利用正域与属性集的单调关系,构造基于属性依赖度和前向搜索策略的快速算法.该算法降低样本比较次数,提高计算效率.实验分析表明该算法的有效性.
Rough set theory is widely used in attribute reduction, the complexity of the algorithm is the main problem of constraint reduction applied to large sample knowledge discovery, especially the reduction problem in neighborhood model.This paper analyzes the mathematical properties of neighborhood rough set model, By using the monotonic relationship between positive domain and attribute set, a fast algorithm based on attribute dependence and forward search strategy is constructed, which reduces the number of sample comparisons and improves the computational efficiency. Experimental results show the effectiveness of the algorithm.