【摘要】 终端信息泄漏是大数据安全的主要问题,大规模高速数据流的潜在异常风险直接影响大数据终端运行状态。为此提出基于Spark计算的大数据终端潜在异常识别方法。分析终端潜在异常数据的噪声影响程度,利用去噪算法对原始终端数据完成去噪预处理。将其输入网络大数据深度挖掘模型中提取潜在异常数据的特征。以Spark计算和自适应快速决策树为基础构建并行性分类模型,将提取到的特征输入至模型,实现大数据终端潜在异常的识别。仿真结果表明,所提方法识别精确度和效率均较高,且具有更大的适应度,说明研究方法的稳定性更优。更多还原

【Abstract】 Terminal information leakage is the main problem of big data security. The potential abnormal risk directly affects the operation state of big data terminals. In this paper,a method of identifying potential anomaly of big data terminals was put forward based on spark computing. At first,the noise influence of the potential abnormal data was analyzed,and then the denoising algorithm was adopted to complete the preprocessing of the original terminal data. After that,the data was input into the deep mining model of network big data for extracting the characteristics of potential abnormal data. Based on spark computing and adaptive fast decision tree,a parallel classification model was constructed. Finally, the extracted features were input into the model to realize the identification of potential anomalies. Simulation results show that the proposed method has higher recognition accuracy and efficiency,as well as bigger adaptability,indicating that the stability of the method is better.更多还原