基于遗传神经网络的可扩展的相似重复记录检测系统设计与仿真  

Design and Simulation of Extensible Approximately Duplicates Detecting System Based on Genetic Neural Networks

作　　者：肖蕾[1] 郭乐江[1] 唐晓[1] 

机构地区：[1]空军雷达学院,武汉430019

出　　处：《电子工程》2011年第1期42-45,共4页

摘　　要：设计实现了一个可扩展的相似重复记录检测系统（Extensible Approximately Duplicates Detecting System，EADDS），并描述了该系统的设计思想、整体结构以及系统特点。该系统包括预处理模块、聚类模块、字段匹配模块和记录匹配模块，支持聚类算法和字段匹配算法的定制扩充。并通过实验对比了几种著名的算法，实验结果表明该系统提高了相似重复记录检测的精确度。An Extensible Approximately Duplicates Detecting System（EADDS） is designed and implemented, and the design, the overall structure and characteristics of the system are described. The system includes pre-processing module, clustering module, field and record match module. It supports customization and expansion of clustering and field match algorithms. Through experiments, it compares several well-known algorithms, and experimental results show that this method improves the accuracy of approximately duplicate records detection.

关 键 词：遗传神经网络 相似重复记录检测系统 聚类算法 字段匹配算法 

分 类 号：TP311.13[自动化与计算机技术—计算机软件与理论]