欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
随着数据信息的指数式增长,使用单节点来进行入侵检测已经不能满足要求。该论文结合Spark和BP神经网络算法的特点,给出了一种并行化的入侵检测方法。该方法主要使用Spark对BP神经网络进行批量训练,通过多次迭代使实验结果达到较高的精度。相关实验表明,该方法在保证精度的同时,有效减少了BP神经网络的训练时间,实现了较好的加速比,有效地提高了入侵检测的执行效率。
英文摘要：
With the exponential growth of data information, the use of single node for intrusion detection has been unable to meet the requirements. Based on the characteristics of Spark and BP neural network, a parallelized intrusion detection method is proposed. This method mainly uses Spark to carry on the batch training to the BP neural network, passing through the iteration to make the experiment result to achieve the high accuracy. Experimental results show that the proposed method can reduce the training time of BP neural network and achieve better speedup ratio, which can improve the efficiency of intrusion detection effectively.
同期刊论文项目
同项目期刊论文
期刊信息
位置：