云端大数据流序列异常挖掘数学建模仿真

Mathematical Modeling and Simulation of Cloud Big Data Stream Sequence Anomaly Mining

目前已有的数据流序列异常挖掘方法忽略了对数据流融合特征的提取,导致传统方法的挖掘时间长,且方法应用性能低。为此提出云端大数据流序列异常挖掘数学建模方法。利用神经网络方法建立数据流的神经网络模型,检测云端数据流信息,结合布谷鸟算法搜索数据流最佳融合特征。基于提取的数据流特征确定模型阈值自适应调整策略、相关系数以及约束条件。依据确立的相关模型指标完成云端大数据流序列异常挖掘模型的构建。实验结果表明,运用上述方法建立异常挖掘模型时,模型的挖掘时长短,且相关系数和召回率指标均较高。

Currently,the existing methods ignore the extraction for fusion features of data flow,resulting in a long mining time and low application performance.Therefore,a mathematical modeling method for mining abnormal sequences in cloud big data flow was put forward.At first,the neural network method was used to build a neural network model of data flow,and thus to detect the information of cloud data flow.Moreover,the best fusion feature of data flow was searched by the cuckoo algorithm.Based on the extracted features of data flow,the adaptive adjustment strategy,correlation coefficient and constraint conditions of the model threshold were determined.According to relevant indicators,the model of mining abnormal sequences in cloud big data flow was constructed.Experimental results show that when using the above method to establish an anomaly mining model,the mining time of the model is long,and the correlation coefficient and recall index are high.