改进量子粒子群算法优化神经网络的数据库重复记录检测

摘要

为了提高数据库重复记录检测效果，提出一种改进量子粒子群优化算法（IQPSO）优化 BP 神经网络的数据库相似重复记录检测模型（IQPSO-BPNN）。首先计算记录字段间的相似度，组成特征向量；然后采用 IQPSO 算法优化 BP 神经网络进行学习，建立最优相似重复记录检测模型，最后通过仿真实验对 IPSO-BPNN 的性能进行测试。实验结果表明，IQPSO-BPNN 大幅度减少了数据库重复记录检测时间，提高了数据库重复记录检测精度。%To improve detection effect of duplicate records in database,we propose a database similar duplicate records detection model IQPSO-BPNN,which uses the improved quantum particle swam optimisation (IQPSO)to optimise BP neural network.First,it calculates the similarity of the corresponding fields between two records and forms the eigenvectors;and then it uses IQPSO to optimise BP neural network to learn,and builds the optimal similar duplicate records detection model;finally,through simulation experiments it tests the performance of IQPSO-BPNN.Experimental results show that the proposed model significantly reduces the detection time and improves the detection accuracy of database duplicate records.