中文摘要：

提出一种考虑语义特征聚类紧密性与分离性特征分析的并行数据库中异常数据优化分类挖掘算法.构建并行数据库异常数据的语义特征分离性度量模型,设计语义映射网络结构,实现对异常数据的噪点初步分离,设计面向语义特征聚类的紧密性与分离性特征提取算法,对含有噪点和野值的并行数据库环境进行干扰抑制,实现对异常数据库优化分类挖掘.仿真结果表明,该算法提高了对并行数据库中异常数据搜索过程中的挖掘查准率,分类挖掘紧密度和准确度较高.

英文摘要：

Considering semantic feature clustering closely and separation of parallel abnormal data in the database optimization classification algorithm is presented in this paper. Constructing parallel database anomaly data semantic feature separability measure model is obained, design semantic mapping network structure to achieve noise of abnormal data preliminary separation, design semantic features clustering compactness and separation algorithm, to contain noise and outliers in the parallel database system for interference suppression, mining classification of abnormal database optimization. Simulation results show that the algorithm improves the parallel database abnormal data search the mining process of precision and classification mining compactness and high accuracy.