A Fast Pruning Algorithm for Neural Network

QIAO Jun-fei; LI Miao; LIU Jiang; BR

您当前的位置：

首页 >

文章列表页 >

A Fast Pruning Algorithm for Neural Network

更新时间：2025-07-16

- A Fast Pruning Algorithm for Neural Network
- Acta Electronica Sinica Vol. 38, Issue 4, Pages: 830-0834(2010)
- 作者机构：
  
  （北京工业大学电子信息与控制工程学院，）,北京,100124
- 作者简介：
- 基金信息：
- DOI：
  CLC：
- Published：2010
- 稿件说明：
移动端阅览
QIAO Jun-fei, LI Miao, LIU Jiang, et al. A Fast Pruning Algorithm for Neural Network[J]. Acta Electronica Sinica, 2010, 38(4): 830-0834.
DOI：

QIAO Jun-fei, LI Miao, LIU Jiang, et al. A Fast Pruning Algorithm for Neural Network[J]. Acta Electronica Sinica, 2010, 38(4): 830-0834. DOI：

摘要

<FONT face=Verdana>通过分析非线性函数各输入参数对函数值的影响，发现当输入参数间无相互作用时，表征其灵敏度的傅立叶振幅主要集中在基频上。基于该发现，提出一种基于扩展傅立叶振幅灵敏度方法的神经网络隐含层神经元快速修剪算法。其实质是通过计算神经网络隐含层各神经元输出的指定基频上的傅立叶振幅，求取隐含层神经元对神经网络输出的灵敏度。根据各神经元的灵敏度，按照一定的准则削减冗余神经元，获得紧凑的神经网络结构。将提出的神经网络结构修剪算法用于污水水质参数化学需氧量(COD)的软测量过程中，实验结果与扩展傅里叶振幅灵敏度算法相比，在修剪效果相同的情况下，其运行时间得到明显减小。

Abstract

<FONT face=Verdana>By analyzing the impact of the input factors on the output value in nonlinear function

it is suggested that when the input factors are independent

the Fourier Amplitudes which showing sensitivity values are relied mainly on the fundamental frequency. As a result

a fast pruning algorithm for the hidden neurons in the neural network is proposed based on the Fourier amplitude sensitivity test method. In essence

the Fourier amplitudes on the assigned frequencies of the hidden layer outputs are computed. Then the sensitivity of each hidden neuron to the neural network output is obtained. Finally

the redundant hidden neurons are pruned according to their sensitivity values to obtain a network with compact structure. The propose method is used in the soft measurement for Chemical Oxygen Demand(COD)

which is a quality parameter of waste water. The experimental result shows that our proposed method is much faster than the Fourier amplitude sensitivity test method. The remaining neurons are the same after pruning for the two methods .

关键词

Keywords

references

Views

1418

下载量

CSCD

Alert me when the article has been cited

提交

Tools

Publicity Resources

Numerical Simulation Study on Thickness Calculation of Dural Hematoma Based on Monte Carlo Method

A Unified Adjoint Network Method of Linear and Nonlinear Circuits

Research on Super-Low Frequency Mechanical Antenna Model and Experimental Study of Magnetic Sensor Coil

Study on High Sensitivity and Low Noise Electro-Optic Terahertz Detector

Topological Analysis of Network Sensitivity and Time Complexity of Algorithm

Related Author

GUO Pan

QIN Rui

LUO Xia

XIANG Yang

LI Yan-yan

LUO Hai-jun

FAN Xin-yan

Zhang Zu-shun

Related Institution

Chongqing National Center for Applied Mathematics

Chongqing Key Laboratory of Optoelectronic Functional Materials, Chongqing Normal University

Box 141,Nanjing University of Post and Telecommunication

College of Information Science and Engineering Southeast University Nanjing Jiangsu China

Aerospace Engineering University Beijing China

⁰