欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
迭代最优化算法是模式识别中重要的无指导学习方法。算法因随机确定k个聚类中心进行初始划分的原因，存在初始聚类中心选择的盲目性、容易陷入局部极值、忽略样本的聚类趋势等缺点。经过对迭代最优化算法的研究与分析，根据样本的聚类趋势，结合邻域思想，改进了聚类中心的选择方法，设计了基于样本邻域概念的迭代最优化算法，算法总的时间代价为O（n）。该算法已应用于基于SNMP协议的网络故障管理中的故障分析，分析结果与实际故障类型基本一致，并为计算机网络故障分析提供了一种可行的分析方法。
英文摘要：
The iterative optimization algorithm is an important method of the unsupervised pattern classification. The center of classes that will be elementary classified in original phases is defined by random method in this algorithm. Because of this reason, the iterative optimization algorithm has some serious defects. The defects were selected samples blindly, presented a local extremum in iterative optimization and didn＇ t pay attention to clustering tendency of samples. By the researching and analy- zing, the newly algorithm, designed iterative optimization algorithm based on neighborhood of samples, according to the con- ception of the clustering tendency and neighborhood of patterns. The time complexity of the newly algorithm was O （n） and n was a number of samples in sets. is Applied this algorithm in the faults analysis in network management based on SNMP protocol. The analysis results are consistent with faults type and it provides a feasible method for network faults analysis.
同期刊论文项目
同项目期刊论文
期刊信息
位置：