论文部分内容阅读
提出一种新的模糊分类模型,该模型利用决策粗糙集指导模糊分类模型结构的设计.首先采用模糊C均值聚类算法对连续属性离散化同时对输入空间进行模糊划分;然后利用两步搜索策略计算离散化决策表的约简,删除冗余的条件属性;从约简后的决策表中提取决策规则,再将决策规则转换成模糊分类规则,从而建立了模糊分类模型.模糊分类模型的规则物理含义明确、形式简化,并且不需要再采用学习算法调整模型的参数.最后利用UCI(university of California irvine)标准数据集与现有的一些分类算法进行了比较,仿真实验结果证明了本文提出的模型是有效的.
This paper proposes a new fuzzy classification model which uses decision-making rough set to guide the design of fuzzy classification model structure.Firstly, fuzzy C-means clustering algorithm is used to discretize the continuous attributes and the input space is divided at the same time. Then, a two-step search strategy The discretization decision table is reduced and the redundant condition attributes are deleted, the decision rules are extracted from the reduced decision table, and then the decision rules are converted into fuzzy classification rules to establish the fuzzy classification model.The rules of the fuzzy classification model The physical meaning is clear, the form is simplified, and no need to use the learning algorithm to adjust the parameters of the model.Finally, using the UCI standard dataset compared with some existing classification algorithms, the simulation results prove that the proposed The model is valid.