东篱科研大数据发现系统（DRDS）

位置：成果数据库 > 期刊 > 期刊详情页

高维数据多级模糊模式识别的分类研究

期刊名称：计算机应用研究
时间：0
页码：4045-4047+4053
语言：中文
分类：TP274[自动化与计算机技术—控制科学与工程;自动化与计算机技术—检测技术与自动化装置]
作者机构：[1]哈尔滨理工大学计算机科学与技术学院,哈尔滨150080, [2]哈尔滨工程大学计算机科学与技术学院,哈尔滨150001
相关基金：国家自然科学基金资助项目（60873019,60673131）;黑龙江省自然科学基金资助项目（F200608）;黑龙江省教育厅海外学人重点科研资助项目（1152hq08）
相关项目：单件复杂产品加工和装配过程综合调度优化算法

关键词：分类, 多级模糊模式识别, 权重, 约简, 高维数据, classification, multiple fuzzy pattern recognition , weight , reduction, high-dimensional data

中文摘要：

通过分析对象属性间的关系，提出了一种基于改进的多级模糊模式识别的分类方法。该方法重点考虑对象属性间影响较大的因素，以此建立影响对象分类的属性之间的简化关系，使分类结果更加合理；针对分类标准为对象属性分类的离散值，存在对象属性值介于中间状态不便分类问题，通过建立属性值所属级别的矩阵来确定属性权重，使分类精确；利用Rough集的特征属性约简算法降低数据集的维数，提高高维数据的分类效率。经实例证明该方法分类准确、效率高。

英文摘要：

By analysing the relation among object attribute, this paper introduced a classification method based on improved multiple fuzzy pattern recognition. This method mainly took the factors into account which were greatly affected by object attribute, so as to find out the concise relation affecting object classification among attribute and make the classification result more rational; as for the question that it was inconvenient for classification owing to betweenness because classification standard was described by discrete value of object cassification, introduced a matrix of attribute value belonging to classification to estimate attribute weight, so it could make the classification more accurate. This study made use of algorithm on the characteristic attribute reduction of rough set in order to reduce dimension of data set and enhance classification efficiency in high-dimensional data. It proves this method makes classification more accurate and more efficient with the example.

同期刊论文项目