网络数据库中隐蔽数据快速挖掘方法研究

摘要

In the framework of large network database, contains a mass of pictures, voice, text, etc. data information, because the difference between the data and disturbance, resulting in treat access the target data of strong concealment, the rapid excavation of hidden data is network database access optimization based. Traditional method uses the fuzzy C means clustering algorithm for data mining, the anti interference of the algorithm is not strong, the classification of dynamic differential data mining performance is not high. A fast data mining algorithm for hidden data in the network database based on the feature point detection of the data time frequency distribution is proposed. Data network data distribution structure model, time series analysis and data signal model is constructed and of FCM clustering preprocessing network database in the data and the output of the clustering of data frequency analysis and feature point detection, to achieve accurate data mining proposed. Simulation results show that the proposed algorithm is of high accuracy, fast convergence and good performance.%在大型网络数据库构架中，包含有海量的图片、声音、文字等数据信息，由于数据之间的差异性较大以及扰动干扰，导致对待访问的目标数据的隐蔽性较强，对隐蔽数据的快速挖掘是实现网络数据库优化访问的基础。传统方法采用模糊C均值聚类算法进行数据挖掘，算法的抗干扰性不强，动态差异性数据的分类挖掘性能不高。提出一种基于数据时频分布特征点检测的网络数据库中隐蔽数据快速挖掘算法。构建网络数据的数据分布结构模型，进行数据时间序列分析和信号模型构建，对网络数据库中的大数据进行FCM聚类预处理，对聚类输出的数据进行时频分析和特征点检测，实现数据准确挖掘。仿真结果表明，采用该算法进行数据挖掘的准确度较高，快速收敛性较好，展示了较好的性能。