混合存储视频数据库中自适应数据挖掘仿真  

Simulation of Adaptive Data Mining Method in Hybrid Storage Video Database

作　　者：刘红敏[1] LIU Hong-min(Guangzhou University Sontian College Guangdong guangzhou 511370,China)

机构地区：[1]广州大学松田学院

出　　处：《计算机仿真》2019年第6期259-262,共4页Computer Simulation

摘　　要：针对传统的数据挖掘方法存在挖掘时间过长,实时响应性较差等问题,提出一种混合存储视频数据库中自适应数据挖掘方法。构建混合存储视频数据存储结构和传输信道模型,根据其模型对混合视频数据传输信道进行多普勒扩展来降低混合视频数据挖掘过程中的通信传输衰减损失,通过级联滤波算法对混合视频数据进行滤波处理,根据其滤波处理后的数据中提取出混合视频数据频谱特征;通过K-means方法对提取出的特征进行分类,根据微粒子群优化算法将其分类结果视为微粒,根据自身飞行经验来不断调整其速度,运用某种规则对调整后的最优速度进行更新,根据其结果实现数据挖掘。实验结果证明,所提方法可以有效减少数据挖掘的时间,提高数据挖掘的实时响应性能。Traditionally, the data mining method leads to long mining time and poor real-time responsiveness. Therefore, an adaptive data mining method in hybrid storage video database is proposed. The model of storage structure and transmission channel of hybrid storage video data was constructed. According to the model, Doppler expansion was performed on the hybrid video data transmission channel to reduce the communication transmission attenuation loss during the hybrid video data mining. The hybrid video data was filtered by the cascaded filter algorithm. According to the filtered data, the spectral features of mixed video data were extracted. K-means method was used to classify extracted features. According to the particle swarm optimization algorithm, the classification results were regarded as particles. Based on its own flying experience, its speed was adjusted constantly. On the basis of some rules, the adjusted optimal speed was updated. Finally, the data mining was achieved based on the results. Simulation results show that the proposed method can effectively reduce the time of data mining and improve the real-time response performance of data mining.

关 键 词：频谱特征 微粒子群算法 最优值 

分 类 号：TP391.41[自动化与计算机技术—计算机应用技术]