摘 要:针对冗余参数占用较多网络存储空间的问题,提出一种基于大数据的深度卷积神经网络冗余参数优化算法。根据大数据的特点来确定冗余参数特点,由此构建基于大数据的深度卷积神经网络模型。计算属性数据集合的冗余参数熵,以熵值为依据压缩冗余参数。通过梯度下降的方式训练深度卷积神经网络参数,确定冗余参数拥塞程度,控制数据冗余拥塞队列,以此完成冗余参数优化。由实验结果可知,该算法所占网络存储空间最小为 2 GB,参数可用效率最高可达 92%,具有良好的优化效果。
关键词:大数据;深度卷积神经网络;冗余参数优化
DOI:10.19850/j.cnki.2096-4706.2022.07.022
中图分类号:TP183 文献标识码:A 文章编号:2096-4706(2022)07-0088-03
Redundant Parameter Optimization Algorithm of Deep Convolution Neural Network Based on Big Data
JIANG Ning
(College of Mathematics and Computer Science, Yan'an University, Yan'an, 716000 China)
Abstract: Aiming at the problem that redundant parameters occupy more network storage space, a deep convolution neural network redundant parameter optimization algorithm based on big data is proposed. The characteristics of redundant parameters are determined according to the characteristics of big data, and a deep convolution neural network model based on big data is constructed. Calculate the redundant parameter entropy of attribute data set, and compress the redundant parameters based on the entropy value. The parameters of deep convolution neural network are trained by gradient descent to determine the congestion degree of redundant parameters and control the data redundant congestion queue, so as to complete the optimization of redundant parameters. The experimental results show that the minimum network storage space occupied by the algorithm is 2 GB, and the maximum parameter availability efficiency can reach 92%, which has a good optimization effect.
Keywords: big data; deep convolution neural network; optimization of redundant parameters
参考文献:
[1] 毛伊敏,张瑞朋,曹文梁 . 基于大数据的并行化深度卷积神经网络优化算法 [J]. 计算机应用研究,2021,38(5):1416-1421.
[2] 谢斌红,钟日新,潘理虎,等 . 结合剪枝与流合并的卷积神经网络加速压缩方法 [J]. 计算机应用,2020,40(3):621-625.
[3] 刘永利,朱亚孟,晁浩 . 多策略 MRFO 算法的卷积神经网络超参数优化 [J]. 北京邮电大学学报,2021,44(6):83-88+95.
[4] 陈靓,钱亚冠,何志强,等 . 深度卷积神经网络的柔性剪枝策略 [J]. 电信科学,2022,38(1):83-94.
[5] 孟丽楠,谢红薇,宁晨,等 . 基于深度卷积神经网络的心音分类算法 [J]. 计算机测量与控制,2021,29(8):211-217+222.
[6] 陈薪羽,刘明哲,任俊,等 . 基于多列卷积神经网络的参数异步更新算法 [J]. 计算机应用,2022,42(2):395-403.
作者简介:姜宁(1983—),女,汉族,陕西临潼人,实验师,硕士,研究方向:实验室管理与维护。