摘要
在不完备信息系统中,可变精度分类关系是限制容差关系的改进形式,但其并未考虑数据集中属性的测试代价。为解决这一问题,提出了基于测试代价敏感的可变精度分类粗糙集模型。进一步地,通过分析传统启发式算法没有考虑测试代价以及回溯算法的时间消耗等因素,提出一种新的属性重要度测量,并在此基础上设计了一种新的启发式算法。通过实验对比分析,说明了新提出算法的有效性。
In an incomplete information system, the precision-variable classification relation is an improvement of the limited tolerance relation. However, the test costs of the data concentration attributes are not taken into account. To solve this problem, a test-cost-sensitive-based precision-variable precision classification rough set is proposed. Furthermore, the traditional heuristic algorithm does not take the importance of the test costs of the attributes into account, and backtracking algorithm is very time-consuming. Therefore, not only was a new importance of the at- tribute proposed, but a new heuristic algorithm was also presented for obtaining reduction with minor test costs. The experimental results show the effectiveness of the new algorithm by comparing it with the other algorithms.
出处
《智能系统学报》
CSCD
北大核心
2014年第2期219-223,共5页
CAAI Transactions on Intelligent Systems
基金
国家自然科学基金资助项目(61100116
61203024)
江苏省自然科学基金资助项目(BK2011492
BK2012700)
江苏省高校自然科学基金资助项目(11KJB520004
13KJB520003)
高维信息智能感知与系统教育部重点实验室(南京理工大学)基金资助项目(30920130122005)
江苏省普通高校研究生科研创新计划项目资助项目(CXLX13_707)
关键词
属性约简
不完备信息系统
测试代价敏感
变精度分类粗糙集
attribute reduction
incomplete information system
test cost sensitive
variable precision classification rough set
