论文部分内容阅读
给出完备决策表和不完备决策表的定义并说明相容关系.给出了相容矩阵及其属性约简的定义,同时也给出差别矩阵及其属性约简的定义,证明了基于相容矩阵的属性约简与关于差别矩阵的属性约简定义是等价的,给出了一个计算条件属性的频率的公式,该公式不必计算差别矩阵,而是直接从决策表中计算出各条件属性在差别矩阵中出现的频率.设计一个快速计算条件属性频率的快速算法,在此基础上,设计了一个高效求基于相容矩阵的属性约简算法,并通过实例对该算法进行了验证.实践证明:算法的复杂度都得以降低,该算法的时间复杂度为O(|C|2|U|),空间复杂度为O(|U|).该方法为计算其他的属性约简算法提供了一条新思路.
The definition of complete decision table and incomplete decision table are given and the compatibility relationship is illustrated. The definition of compatible matrix and its attribute reduction is given. The definition of discernibility matrix and its attribute reduction is also given. The attribute reduction of the capacity matrix is equivalent to the definition of the attribute reduction of the discernibility matrix, and gives a formula for calculating the frequency of the condition attributes. The formula does not need to calculate the difference matrix, but calculates the conditions directly from the decision table The frequency of the attribute in the discernibility matrix.A fast algorithm of calculating the frequency of the conditional attribute is designed rapidly, and then an attribute reduction algorithm based on the consistency matrix is designed and validated by an example. It has been proved by practice that the complexity of the algorithm is reduced, the time complexity of the algorithm is O (| C | 2 | U |) and the space complexity is O (| U |). This method is to calculate other attribute reduction algorithms Provide a new idea.