论文部分内容阅读
为降低经典信息熵属性约简算法的时间复杂度,在论证信息熵属性约简与论域划分细化约简等价的基础上,提出将蚁群并行优化处理机制引入划分细化约简过程,在蚁群搜索过程中使用体现属性约简特点的状态转移和信息素更新策略.通过复杂性分析和实例验证,新算法可有效避免蚁群搜索的盲目性,并在较小迭代规模下快速获得约简集,更适于大容量数据表的处理.
In order to reduce the time complexity of attribute reduction algorithm of classical information entropy, based on the argument that attribute reduction of information entropy and deduction equivalence of domain reduction, the mechanism of parallel optimization of ant colony is introduced into the process of partitioning reduction , The state transition and pheromone updating strategy which reflect the characteristics of attribute reduction are used in the ant colony searching process.The new algorithm can effectively avoid the blindness of ant colony searching through the complexity analysis and the example verification and can be quickly Get reduction set, more suitable for large-capacity data sheet processing.