直觉模糊软集从参数化和程度化两个方面来表述对象,这种表述可以更全面地刻画对象的特征。直觉模糊软集的正规化运算是在保持直觉模糊软集的表达能力的同时,降低其计算复杂度的一种重要手段。本文通过引入犹豫度偏好参数,提出直觉模糊...直觉模糊软集从参数化和程度化两个方面来表述对象,这种表述可以更全面地刻画对象的特征。直觉模糊软集的正规化运算是在保持直觉模糊软集的表达能力的同时,降低其计算复杂度的一种重要手段。本文通过引入犹豫度偏好参数,提出直觉模糊软集正规化运算的一种拓展模型。首先,利用犹豫度偏好参数将犹豫度按照偏好重新分配给隶属函数和非隶属函数,使二者之和为1,得到一种新的正规化运算。其次,讨论这种正规化运算的基本性质。最后,通过一个实际问题验证基于直觉模糊软集正规化运算的模糊决策方法具有合理性和有效性。Intuitionistic fuzzy soft sets characterize objects from the perspective of parameterization and degree, which can describe objects more comprehensively. The normalization operation of intuitionistic fuzzy soft sets is an important means to reduce the complexity of the computation in the process of application while maintaining the expression ability of intuitionistic fuzzy soft sets. In this paper, an improved model of normalization operation of intuitionistic fuzzy soft sets is proposed. Firstly, the hesitation degree is redistributed to membership function and non-membership function according to the preference of decision maker, making the sum of membership value and non-membership value is 1, and a new normalization operation is obtained. Secondly, some elemental properties of this new necessity operation are discussed. Finally, a practical example is used to verify the rationality and effectiveness of the fuzzy decision-making method based on the normalization operation of intuitionistic fuzzy soft sets.展开更多
针对不平衡数据集上进行文本分类,传统的特征选择方法容易导致分类器倾向于大类而忽视小类,提出一种新的特征选择方法 IPR(integrated probability ratio)。该方法综合考虑特征在正类和负类中的分布性质,结合四种衡量特征类别相关性的...针对不平衡数据集上进行文本分类,传统的特征选择方法容易导致分类器倾向于大类而忽视小类,提出一种新的特征选择方法 IPR(integrated probability ratio)。该方法综合考虑特征在正类和负类中的分布性质,结合四种衡量特征类别相关性的指标对特征词进行评分,能够更好地解决传统特征选择方法在不平衡数据集上的不适应性,在不降低大类分类性能的同时提高了小类的识别率。实验结果表明,该方法有效可行。展开更多
文摘直觉模糊软集从参数化和程度化两个方面来表述对象,这种表述可以更全面地刻画对象的特征。直觉模糊软集的正规化运算是在保持直觉模糊软集的表达能力的同时,降低其计算复杂度的一种重要手段。本文通过引入犹豫度偏好参数,提出直觉模糊软集正规化运算的一种拓展模型。首先,利用犹豫度偏好参数将犹豫度按照偏好重新分配给隶属函数和非隶属函数,使二者之和为1,得到一种新的正规化运算。其次,讨论这种正规化运算的基本性质。最后,通过一个实际问题验证基于直觉模糊软集正规化运算的模糊决策方法具有合理性和有效性。Intuitionistic fuzzy soft sets characterize objects from the perspective of parameterization and degree, which can describe objects more comprehensively. The normalization operation of intuitionistic fuzzy soft sets is an important means to reduce the complexity of the computation in the process of application while maintaining the expression ability of intuitionistic fuzzy soft sets. In this paper, an improved model of normalization operation of intuitionistic fuzzy soft sets is proposed. Firstly, the hesitation degree is redistributed to membership function and non-membership function according to the preference of decision maker, making the sum of membership value and non-membership value is 1, and a new normalization operation is obtained. Secondly, some elemental properties of this new necessity operation are discussed. Finally, a practical example is used to verify the rationality and effectiveness of the fuzzy decision-making method based on the normalization operation of intuitionistic fuzzy soft sets.
文摘针对不平衡数据集上进行文本分类,传统的特征选择方法容易导致分类器倾向于大类而忽视小类,提出一种新的特征选择方法 IPR(integrated probability ratio)。该方法综合考虑特征在正类和负类中的分布性质,结合四种衡量特征类别相关性的指标对特征词进行评分,能够更好地解决传统特征选择方法在不平衡数据集上的不适应性,在不降低大类分类性能的同时提高了小类的识别率。实验结果表明,该方法有效可行。