2. 河北科技大学 理学院, 河北 石家庄 050024
2. School of Sciences, Hebei University of Science and Technology, Shijiazhuang 050024, China
模糊集概念[1]是由Zadeh教授在1965年提出的。它解决了对不确定性概念的描述性问题,使得模糊数学在理论和应用方面的研究迅速发展起来,并取得了丰富的研究成果。1982年由波兰数学家Pawlak提出的粗糙集理论一直被认为是处理信息系统和知识发现方面问题的重要工具[2]。在信息系统或者决策表的研究方面,集合近似的定义方式作为一个重要的问题,一直受到研究者的广泛关注。在单个粒度的粗糙集模型下,一个关系产生的一组上下近似常用来刻画一个目标概念的特征。然而在实际生活或者应用中,由于用户需求的不同以及解决问题最终目标的不同,用多个关系来刻画一个目标概念往往更加贴合实际。基于此,钱宇华等[3-6]提出了多粒度粗糙集模型,使得粗糙集理论在解决实际问题方面的应用更加广泛[7-9]。在多粒度粗糙集模型中,乐观多粒度和悲观多粒度是两个不同的基本研究方法。
此外,研究对象很可能来自于不同论域,因而,两个或者多个论域对真实世界的描述更具有一般性,对进一步研究信息表的规则提取具有积极的作用。因此双论域粗糙集模型[10-12]具有很强的研究价值以及实用价值。
在粗糙集理论[13-15]中,上下近似算子是一对基本的概念,从经典的理论意义和直观理解上,它们之间存在着包含关系。但是,当把它们推广到双论域上多粒度粗糙集模型[16-18]时,集合的上下近似并不一定存在着包含关系。本文将就这一问题展开讨论。
1 多粒度模糊粗糙集实际生活中的数据集大部分都是连续的,然而粗糙集研究的一般是离散型数据,利用模糊粗糙集理论就可以解决这一矛盾。对对象集的划分一直是粗糙集研究的重要基础,多粒度粗糙集就是在满足多个关系的情况下对对象集的划分,将这些理论推广到双论域上,使得理论更具有推广性。多粒度模糊粗糙集是将模糊粗糙集与多粒度粗糙集两种理论相结合研究。
定义1[1] 论域
定义2[19] 设
$\underline {\sum\limits_{t = 1}^m {R_t^O} } (X){\rm{( }}x{\rm{)}} = \vee _{t = 1}^m{ \wedge _{z \in U}}[(1 - {R_t}(x, z)) \vee X(z)], \forall x, z \in U$ |
$\overline {\sum\limits_{t = 1}^m {R_t^O} } (X) = {\rm{\sim }}\underline {\sum\limits_{t = 1}^m {R_t^O} } {\rm{(\sim }}X)$ |
$\underline {\sum\limits_{t = 1}^m {R_t^P} } (X){\rm{(}}x{\rm{)}} = \wedge _{t = 1}^m{ \wedge _{z \in U}}[(1 - {R_t}(x, z)) \vee X(z)], \forall x, z \in U$ |
$ \overline {\sum\limits_{t = 1}^m {R_t^P} (} X) = \sim\underline {\sum\limits_{t = 1}^m {R_t^P} } \left( {\sim X} \right) $ |
定义3[20] 设U、V是两个非空有限集合,
定义4[21] 设有序三元组
$ \underline {\Re _{\sum\limits_{t = 1}^m {{R_i}} }^O} (A)(x) = \vee _{t = 1}^m{ \wedge _{y \in V}}[(1 - {R_t}(x, y)) \vee A(y)], \forall x \in U, y \in V $ |
$ \overline {\Re _{\sum\limits_{t = 1}^m {{R_t}} }^O} (A)(x) = \wedge _{t = 1}^m{ \vee _{y \in V}}{\rm{[}}{R_t}(x, y) \wedge A(y){\rm{]}}, \forall x \in U, y \in V $ |
$ \underline {\Re _{\sum\limits_{t = 1}^m {{R_t}} }^P} (A)(x) = \wedge _{t = 1}^m{ \wedge _{y \in V}}[(1 - {R_t}(x, y)) \vee A(y)], \forall x \in U, y \in V $ |
$ \overline {\Re _{\sum\limits_{t = 1}^m {{R_t}} }^P} (A)(x) = \vee _{t = 1}^m{ \vee _{y \in V}}{\rm{[}}{R_t}(x, y) \wedge A(y){\rm{]}}, \forall x \in U, y \in V $ |
当论域
例1 在医疗诊断中,设U={病毒性发热,痢疾,伤寒}=
${{ R}_1} = \left[ {\begin{array}{*{20}{c}} {0.2}&{0.5}&{0.2} \\ {0.1}&{0.3}&{0.7} \\ {0.6}&{0.3}&{0.4} \end{array}} \right]$ |
${{ R}_2} = \left[ {\begin{array}{*{20}{c}} {0.2}&{0.5}&{0.3} \\ {0.4}&{0.4}&{0.3} \\ {0.3}&{0.6}&{0.2} \end{array}} \right]$ |
${{ R}_3} = \left[ {\begin{array}{*{20}{c}} {0.4}&{0.3}&{0.3} \\ {0.5}&{0.5}&{0.4} \\ {0.3}&{0.8}&{0.4} \end{array}} \right]$ |
病人对病情的描述
$\begin{aligned} \overline {\Re _{\sum\limits_{t = 1}^m {{R_t}} }^P} (A)({x_1}) = & [(0.2 \wedge 0.4) \vee (0.5 \wedge 0.6) \vee\\[-10pt] & (0.2 \wedge 0.9)] \vee[(0.2 \wedge 0.4) \vee\\ & (0.5 \wedge 0.6) \vee(0.3 \wedge 0.9)] \vee\\ & [(0.4 \wedge 0.4) \vee (0.3 \wedge 0.6) \vee\\ & (0.3 \wedge 0.9)] = 0.5 \end{aligned} $ |
$\begin{aligned} \underline {\Re _{\sum\limits_{t = 1}^m {{R_t}} }^P} (A)({x_1}) = & ([(1 - 0.2) \vee 0.4] \wedge [(1 - 0.5) \vee 0.6] \wedge\\[-10pt] & [(1 - 0.2) \vee 0.9]) \wedge([(1 - 0.2) \vee 0.4] \wedge\\ & [(1 - 0.5) \vee 0.6] \wedge [(1 - 0.3) \vee 0.9]) \wedge\\ & ([(1 - 0.4) \vee 0.4] \wedge [(1 - 0.3) \vee 0.6] \wedge\\ & [(1 - 0.3) \vee 0.9]) = 0.6 \end{aligned} $ |
同理,可以计算
$\overline {\Re _{\sum\limits_{t = 1}^m {{R_t}} }^P} (A){\rm{ = }}\frac{{0.5}}{{{x_1}}} + \frac{{0.7}}{{{x_2}}} + \frac{{0.6}}{{{x_3}}}$ |
$\underline {\Re _{\sum\limits_{t = 1}^m {{R_t}} }^P} (A){\rm{ = }}\frac{{0.6}}{{{x_1}}} + \frac{{0.5}}{{{x_2}}} + \frac{{0.4}}{{{x_3}}}$ |
另外,可以经过计算得到:
$\begin{aligned} \overline {\Re _{\sum\limits_{t = 1}^m {{R_t}} }^O} (A)({x_1}) = & [(0.2 \wedge 0.4) \vee (0.5 \wedge 0.6) \vee\\[-10pt] & (0.2 \wedge 0.9)] \wedge[(0.2 \wedge 0.4) \vee\\ & (0.5 \wedge 0.6) \vee(0.3 \wedge 0.9)]\wedge\\ & [(0.4 \wedge 0.4) \vee (0.3 \wedge 0.6)\vee\\ & (0.3 \wedge 0.9)] = 0.4 \end{aligned} $ |
$\begin{aligned} \underline {\Re _{\sum\limits_{t = 1}^m {{R_t}} }^O} (A)({x_1}) = & ([(1 - 0.2) \vee 0.4] \wedge [(1 - 0.5) \vee 0.6] \hfill \wedge\\[-12pt] & [(1 - 0.2) \vee 0.9]) \vee ([(1 - 0.2) \vee 0.4] \wedge \\ & [(1 - 0.5) \vee 0.6] \wedge[(1 - 0.3) \vee 0.9]) \vee\\ & ([(1 - 0.4) \vee 0.4] \wedge [(1 - 0.3) \vee 0.6] \wedge \hfill \\ & [(1 - 0.3) \vee 0.9]) = 0.6 \end{aligned} $ |
同样,可以计算A的乐观上下近似分别为
$\overline {\Re _{\sum\limits_{t = 1}^m {{R_t}} }^O} (A){\rm{ = }}\frac{{0.4}}{{{x_1}}} + \frac{{0.4}}{{{x_2}}} + \frac{{0.4}}{{{x_3}}}$ |
$\underline {\Re _{\sum\limits_{t = 1}^m {{R_t}} }^O} (A){\rm{ = }}\frac{{0.6}}{{{x_1}}} + \frac{{0.7}}{{{x_2}}} + \frac{{0.6}}{{{x_3}}}$ |
其中,
由计算结果可知,双论域上集合A的乐观与悲观上下近似并不存在包含关系,例如
由上一章可以看到,将单论域上集合的上下近似定义推广到双论域时,其上下近似不一定具有包含关系。下面给出双论域上给定集合
命题1 当
证明 不妨设:
${{ R}_1} = \left[ {\begin{array}{*{20}{c}} {a_{11}^1}&{a_{12}^1}&{a_{13}^1} \\ {a_{21}^1}&{a_{22}^1}&{a_{23}^1} \\ {a_{31}^1}&{a_{32}^1}&{a_{33}^1} \end{array}} \right]$ |
${{ R}_2} = \left[ {\begin{array}{*{20}{c}} {a_{11}^2}&{a_{12}^2}&{a_{13}^2} \\ {a_{21}^2}&{a_{22}^2}&{a_{23}^2} \\ {a_{31}^2}&{a_{32}^2}&{a_{33}^2} \end{array}} \right]$ |
${{ R}_3} = \left[ {\begin{array}{*{20}{c}} {a_{11}^3}&{a_{12}^3}&{a_{13}^3} \\ {a_{21}^3}&{a_{22}^3}&{a_{23}^3} \\ {a_{31}^3}&{a_{32}^3}&{a_{33}^3} \end{array}} \right]$ |
式中:
首先证明悲观近似算子的情况。
$\begin{aligned} \underline {\Re _{\sum\limits_{t = 1}^m {{R_t}} }^P} (A)({x_1}) = & ([(1 - a_{11}^1) \vee {d_1}] \wedge [(1 - a_{12}^1) \vee {d_2}] \wedge\\[-10pt] & [(1 - a_{1{\rm{3}}}^1) \vee {d_{\rm{3}}}])\wedge([(1 - a_{11}^{\rm{2}}) \vee {d_1}] \wedge\\ & [(1 - a_{12}^{\rm{2}}) \vee {d_2}] \wedge[(1 - a_{1{\rm{3}}}^{\rm{2}}) \vee {d_{\rm{3}}}])\wedge \\ & ([(1 - a_{11}^{\rm{3}}) \vee {d_1}] \wedge [(1 - a_{12}^{\rm{3}}) \vee {d_2}] \wedge\\ & [(1 - a_{1{\rm{3}}}^{\rm{3}}) \vee {d_{\rm{3}}}]) \end{aligned} $ |
$\begin{aligned} \overline {\Re _{\sum\limits_{t = 1}^m {{R_t}} }^P} (A)({x_1}) = & [(a_{11}^1 \wedge {d_1}) \vee (a_{12}^1 \wedge {d_2}) \vee (a_{1{\rm{3}}}^1 \wedge {d_{\rm{3}}})] \vee \\ & [(a_{11}^{\rm{2}} \wedge {d_1}) \vee (a_{12}^{\rm{2}} \wedge {d_2}) \vee (a_{1{\rm{3}}}^{\rm{2}} \wedge {d_{\rm{3}}})] \vee \\ & [(a_{11}^{\rm{3}} \wedge {d_1}) \vee (a_{12}^{\rm{3}} \wedge {d_2}) \vee (a_{1{\rm{3}}}^{\rm{3}} \wedge {d_{\rm{3}}})]& \end{aligned} $ |
并且对于任意两个数
$\begin{gathered} p \vee q = \frac{{p + q}}{2} + \frac{{|p - q|}}{2} \hfill \\ p \wedge q = \frac{{p + q}}{2} - \frac{{|p - q|}}{2} \hfill \\ \end{gathered} $ |
那么,
由定义4可以看到,双论域上的多粒度近似算子同
$\begin{array}{c} [(1 - a_{ij}^t) \vee {d_j}] - [a_{ij}^t \wedge {d_j}] = \\ \displaystyle\frac{{1 - a_{ij}^t + {d_j}}}{2} + \frac{{|1 - a_{ij}^t - {d_j}|}}{2} - \left[\frac{{a_{ij}^t + {d_j}}}{2} - \frac{{|a_{ij}^t - {d_j}|}}{2}\right] = \\ \displaystyle\frac{{1 - a_{ij}^t + {d_j}}}{2} + \frac{{|1 - a_{ij}^t - {d_j}|}}{2} - \frac{{(a_{ij}^t + {d_j})}}{2} + \frac{{|a_{ij}^t - {d_j}|}}{2} = \\ \displaystyle\frac{{1 - 2a_{ij}^t}}{2} + \frac{{|1 - a_{ij}^t - {d_j}| + |a_{ij}^t - {d_j}|}}{2} \geqslant \\ \displaystyle\frac{{1 - 2a_{ij}^t}}{2} + \frac{{|1 - a_{ij}^t - {d_j} + a_{ij}^t - {d_j}|}}{2} =\frac{{1 - 2a_{ij}^t}}{2} + \frac{{|1 - 2{d_j}|}}{2}\geqslant \\ 1 - (a_{ij}^t + {d_j}) \end{array} $ |
$ \begin{array}{c} [(1 - a_{ij}^t) \vee {d_j}] - [a_{ij}^{t'} \wedge {d_j}] =\\ \displaystyle \frac{{1 - a_{ij}^t + {d_j}}}{2} + \frac{{|1 - a_{ij}^t - {d_j}|}}{2} - \left[\frac{{a_{ij}^{t'} + {d_j}}}{2} - \frac{{|a_{ij}^{t'} - {d_j}|}}{2}\right] = \\ \displaystyle \frac{{1 - a_{ij}^t + {d_j}}}{2} + \frac{{|1 - a_{ij}^t - {d_j}|}}{2} - \frac{{(a_{ij}^{t'} + {d_j})}}{2} + \frac{{|a_{ij}^{t'} - {d_j}|}}{2} = \\ \displaystyle\frac{{1 - (a_{ij}^t + a_{ij}^{t'})}}{2} + \frac{{|1 - a_{ij}^t - {d_j}| + |a_{ij}^{t'} - {d_j}|}}{2} \geqslant \\ \displaystyle\frac{{1 - (a_{ij}^t + a_{ij}^{t'})}}{2} + \frac{{|1 - a_{ij}^t - {d_j}| + a_{ij}^{t'} - {d_j}}}{2} = \\ \displaystyle\frac{{1 - a_{ij}^t - {d_j}}}{2} + \frac{{|1 - a_{ij}^t - {d_j}|}}{2} \geqslant 0 \end{array} $ |
$ \begin{array}{c} [(1 - a_{ij}^t) \vee {d_j}] - [a_{hk}^t \wedge {d_k}] = \\ \displaystyle\frac{{1 - a_{ij}^t + {d_j}}}{2} + \frac{{|1 - a_{ij}^t - {d_j}|}}{2} - \left[\frac{{a_{hk}^t + {d_k}}}{2} - \frac{{|a_{hk}^t - {d_k}|}}{2}\right] = \\ \displaystyle\frac{{1 - a_{ij}^t + {d_j}}}{2} + \frac{{|1 - a_{ij}^t - {d_j}|}}{2} - \frac{{(a_{hk}^t + {d_k})}}{2} + \frac{{|a_{hk}^t - {d_k}|}}{2} = \\ \displaystyle \frac{{1 - a_{ij}^t - a_{hk}^t + {d_j} - {d_k}}}{2} + \frac{{|1 - a_{ij}^t - {d_j}| + |a_{hk}^t - {d_k}|}}{2} \geqslant \\ \displaystyle\frac{{1 - a_{ij}^t - a_{hk}^t + {d_j} - {d_k}}}{2} + \frac{{|1 - a_{ij}^t| - {d_j} + |a_{hk}^t - {d_k}|}}{2} = \\ \displaystyle\frac{{1 - a_{ij}^t - a_{hk}^t - {d_k}}}{2} + \frac{{|1 - a_{ij}^t| + |a_{hk}^t - {d_k}|}}{2} \geqslant \\ \displaystyle\frac{{1 - a_{ij}^t - a_{hk}^t - {d_k}}}{2} + \frac{{|1 - a_{ij}^t| + a_{hk}^t - {d_k}}}{2} = \\ \displaystyle \frac{{1 - a_{ij}^t - {d_k}}}{2} + \frac{{|1 - a_{ij}^t| - {d_k}}}{2} = 1 - (a_{ij}^t + {d_k}) \end{array} $ |
由上面3个证明过程可以推出,在悲观的情况下,当
同理可证:乐观的情况下,上下近似算子中的基本元素并没有发生改变,所以仍满足下近似中的任意元素都大于上近似中的任意元素,故最终求得的包含度仍具有相同的大小关系。
一般情况,当论域基数变大且给定粒度的个数推广至
命题2 当
证明 设
${{ R}_t} = \left[ {\begin{array}{*{20}{c}} {a_{11}^t}& {a_{12}^t}\cdots &{a_{1l}^t} \\ {a_{21}^t}&{a_{22}^t}\cdots &{a_{2l}^t} \\ \vdots & \vdots& \vdots \\ {a_{n1}^t}& {a_{n2}^t} &{a_{nl}^t} \end{array}} \right]$ |
其中
$\begin{aligned} \underline {\Re _{\sum\limits_{t = 1}^m {{R_t}} }^P} (A)({x_1}) = & ([(1 - a_{11}^1) \vee {d_1}] \wedge [(1 - a_{12}^1) \vee {d_2}] \cdots \wedge \\[-10pt] & [(1 - a_{1l}^1) \vee {d_l}]) \cdots \wedge([(1 - a_{11}^m) \vee {d_1}] \wedge\\ & [(1 - a_{12}^m) \vee {d_2}] \cdots \wedge [(1 - a_{1l}^m) \vee {d_l}]) \end{aligned} $ |
$\begin{aligned} \overline {\Re _{\sum\limits_{t = 1}^m {{R_t}} }^P} (A)({x_1}) = & [(a_{11}^1 \wedge {d_1}) \vee (a_{12}^1 \wedge {d_2}) \cdots \vee \\[-10pt] & (a_{1l}^1 \wedge {d_l})] \cdots \vee [(a_{11}^m \wedge {d_1}) \vee\\ & (a_{12}^m \wedge {d_2}) \cdots \vee (a_{1l}^m \wedge {d_l})] \end{aligned} $ |
类比命题1的证明过程并改变索引集的取值范围后,可得结论:当
同理,对A中的对象
对于乐观的情况,根据命题1同理可证其上下近在满足上述条件时仍具有包含关系。由此可以得到结论:当
在本章研究基础上,对于双论域上的多粒度粗糙集上下近似不具备包含关系的,将给出标准化的方法,使之转化为具有包含关系。
3 标准化方法由第2章的证明可知,要使双论域上的多粒度粗糙集
定义5 在多粒度空间
由上述定义可知
例2 续例1:
${l_{11}} = 0.2 + 0.1 + 0.6 = 0.9$ |
$I_{11}^{1 + } = \frac{{I_{{\rm{11}}}^{\rm{1}}}}{{{l_{11}}}} = \frac{{0.2}}{{2 \times 0.9}} = \frac{1}{9}$ |
对
${ R}_1^ + = \left[ {\begin{array}{*{20}{c}} {\displaystyle\frac{1}{9}}&{\displaystyle\frac{5}{{22}}}&{\displaystyle\frac{1}{{13}}} \\ {\displaystyle\frac{1}{{18}}}&{\displaystyle\frac{3}{{22}}}&{\displaystyle\frac{7}{{26}}} \\ {\displaystyle\frac{1}{3}}&{\displaystyle\frac{3}{{22}}}&{\displaystyle\frac{2}{{13}}} \end{array}} \right]$ |
${ R}_2^ + = \left[ {\begin{array}{*{20}{c}} {\displaystyle\frac{1}{9}}&{\displaystyle\frac{1}{6}}&{\displaystyle\frac{3}{{16}}} \\ {\displaystyle\frac{2}{9}}&{\displaystyle\frac{2}{{15}}}&{\displaystyle\frac{3}{{16}}} \\ {\displaystyle\frac{1}{6}}&{\displaystyle\frac{1}{5}}&{\displaystyle\frac{1}{8}} \end{array}} \right]$ |
${ R}_3^ + = \left[ {\begin{array}{*{20}{c}} {\displaystyle\frac{1}{6}}&{\displaystyle\frac{3}{{32}}}&{\displaystyle\frac{3}{{22}}} \\ {\displaystyle\frac{5}{{24}}}&{\displaystyle\frac{5}{{32}}}&{\displaystyle\frac{2}{{11}}} \\ {\displaystyle\frac{1}{8}}&{\displaystyle\frac{1}{4}}&{\displaystyle\frac{2}{{11}}} \end{array}} \right]$ |
A进行标准化后为
${A^ + } = \displaystyle\frac{{\displaystyle\frac{2}{{19}}}}{{{x_1}}} + \displaystyle\frac{{\displaystyle\frac{3}{{19}}}}{{{x_2}}} + \displaystyle\frac{{\displaystyle\frac{9}{{38}}}}{{{x_3}}}$ |
则由双论域上多粒度粗糙集(乐观、悲观)上下近似的定义,可以求得
$\underline {\Re _{\sum\limits_{t = 1}^m {R_t^ + } }^P} ({A^ + }){\rm{ = }}\displaystyle\frac{{\displaystyle\frac{{17}}{{22}}}}{{{x_1}}} + \displaystyle\frac{{\displaystyle\frac{{19}}{{26}}}}{{{x_2}}} + \displaystyle\frac{{\displaystyle\frac{2}{3}}}{{{x_3}}}$ |
$\overline {\Re _{\sum\limits_{t = 1}^m {R_t^ + } }^P} ({A^ + }){\rm{ = }}\displaystyle\frac{{\displaystyle\frac{3}{{16}}}}{{{x_1}}} +\displaystyle \frac{{\displaystyle\frac{9}{{38}}}}{{{x_2}}} + \displaystyle\frac{{\displaystyle\frac{2}{{11}}}}{{{x_3}}}$ |
$\underline {\Re _{\sum\limits_{t = 1}^m {R_t^ + } }^O} ({A^ + }){\rm{ = }}\displaystyle\frac{{\displaystyle\frac{5}{6}}}{{{x_1}}} + \displaystyle\frac{{\displaystyle\frac{{19}}{{24}}}}{{{x_2}}} + \displaystyle\frac{{\displaystyle\frac{4}{5}}}{{{x_3}}}$ |
$\overline {\Re _{\sum\limits_{t = 1}^m {R_t^ + } }^O} ({A^ + }){\rm{ = }}\displaystyle\frac{{\displaystyle\frac{3}{{22}}}}{{{x_1}}} + \displaystyle\frac{{\displaystyle\frac{2}{{11}}}}{{{x_2}}} + \displaystyle\frac{{\displaystyle\frac{2}{{13}}}}{{{x_3}}}$ |
由计算结果可知:双论域上的多粒度粗糙集
本文证明了双论域下多粒度模糊粗糙集上下近似具有包含关系的一个充分条件为
为了表述和计算的方便,在进行标准化的时候限定所有
[1] | ZADEH L A. Fuzzy sets[J]. Information and control, 1965, 8(3): 338-353. DOI:10.1016/S0019-9958(65)90241-X (0) |
[2] | PAWLAK Z. Rough sets[J]. International journal of computer and information sciences, 1982, 11(5): 341-356. DOI:10.1007/BF01001956 (0) |
[3] | QIAN Yuhua, LIANG Jiye, DANG Chuangyin. Incomplete multigranulation rough set[J]. IEEE transactions on systems, man, and cybernetics-part a: systems and humans, 2010, 40(2): 420-431. DOI:10.1109/TSMCA.2009.2035436 (0) |
[4] | QIAN Yuhua, LI Shunyong, LIANG Jiye, et al. Pessimistic rough set based decisions: a multigranulation fusion strate-gy[J]. Information sciences, 2014, 264: 196-210. DOI:10.1016/j.ins.2013.12.014 (0) |
[5] | QIAN Yuhua, ZHANG Hu, SANG Yanli, et al. Multigranu-lation decision-theoretic rough sets[J]. International journal of approximate reasoning, 2014, 55(1): 225-237. DOI:10.1016/j.ijar.2013.03.004 (0) |
[6] | QIAN Yuhua, LIANG Jiye, YAO Yiyu, et al. MGRS: a multigranulation rough set[J]. Information sciences, 2010, 180(6): 949-970. DOI:10.1016/j.ins.2009.11.023 (0) |
[7] | XU Weihua, LI Wentao, ZHANG Xiantao. Generalized multigranulation rough sets and optimal granularity selec-tion[J]. Granular computing, 2017, 2(4): 271-288. DOI:10.1007/s41066-017-0042-9 (0) |
[8] | TAN Anhui, WU Weizhi, TAO Yuzhi. On the belief struc-tures and reductions of multigranulation spaces with deci-sions[J]. International journal of approximate reasoning, 2017, 88: 39-52. DOI:10.1016/j.ijar.2017.05.005 (0) |
[9] | QIAN Yuhua, LIANG Xinyan, LIN Guoping, et al. Local multigranulation decision-theoretic rough sets[J]. Interna-tional journal of approximate reasoning, 2017, 82: 119-137. DOI:10.1016/j.ijar.2016.12.008 (0) |
[10] |
邱雅竹, 付蓉. 两个论域上的粗糙集模型及其应用[J]. 四川师范大学学报 (自然科学版), 2005, 28(1): 15-18. QIU Yazhu, FU Rong. Rough set model over two universes and application[J]. Journal of Sichuan normal university (natural science), 2005, 28(1): 15-18. (0) |
[11] | SUN Bingzhen, MA Weimin. Fuzzy rough set model on two different universes and its application[J]. Applied mathematical modelling, 2011, 35(4): 1798-1809. DOI:10.1016/j.apm.2010.10.010 (0) |
[12] | ZHANG Chao, LI Deyu, LIANG Jiye. Hesitant fuzzy linguistic rough set over two universes model and its ap-plications[J]. International journal of machine learning and cybernetics, 2018, 9(4): 577-588. DOI:10.1007/s13042-016-0541-z (0) |
[13] | HUANG Bing, GUO Chunxiang, LI Huaxiong, et al. An intuitionistic fuzzy graded covering rough set[J]. Knowledge-based systems, 2016, 107: 155-178. DOI:10.1016/j.knosys.2016.06.006 (0) |
[14] | 张文修, 梁怡, 吴伟志. 信息系统与知识发现[M]. 北京: 科学出版社, 2003: 1–259. (0) |
[15] | 张文修, 梁怡, 徐萍. 基于包含度的不确定推理[M]. 北京: 清华大学出版社, 2007: 1–310. (0) |
[16] |
梁美社, 米据生, 赵天娜. 广义优势多粒度直觉模糊粗糙集及规则获取[J]. 智能系统学报, 2017, 12(6): 883-888. LIANG Meishe, MI Jusheng, ZHAO Tianna. Generalized dominance-based multi-granularity intuitionistic fuzzy rough set and acquisition of decision rules[J]. CAAI transactions on intelligent systems, 2017, 12(6): 883-888. (0) |
[17] | ZHANG Chao, Li Deyu, SANGAIAH A K, et al. Merger and acquisition target selection based on interval neutro-sophic multigranulation rough sets over two universes[J]. Symmetry, 2017, 9(7): 126. DOI:10.3390/sym9070126 (0) |
[18] | ZHANG Chao, LI Deyu, YAN Yan. A dual hesitant fuzzy multigranulation rough set over two-universe model for medical diagnoses[J]. Computational and mathematical methods in medicine, 2015, 5: 1-12. (0) |
[19] | XU Weihua, WANG Qiaorong, ZHANG Xiantao. Mul-ti-granulation fuzzy rough sets in a fuzzy tolerance ap-proximation space[J]. International journal of fuzzy systems, 2011, 13(4): 246-259. (0) |
[20] | SUN Bingzhen, MA Weimin. Multigranulation rough set theory over two universes[J]. Journal of intelligent and fuzzy systems, 2015, 28(3): 1251-1269. (0) |
[21] | SUN Bingzhen, MA Weimin, QIAN Yuhua. Multigranula-tion fuzzy rough set over two universes and its application to decision making[J]. Knowledge-based systems, 2017, 123: 61-74. DOI:10.1016/j.knosys.2017.01.036 (0) |