【BP分类】基于反向传播的多层感知器神经网络附matlab代码

✅作者简介：热爱科研的Matlab仿真开发者，修心和技术同步精进，matlab项目合作可私信。

个人主页：Matlab科研工作室

个人信条：格物致知。

⛄ 内容介绍

【BP分类】基于反向传播的多层感知器神经网络附matlab代码

⛄ 部分代码

%% Backpropagation for Multi Layer Perceptron Neural Networks %%

%   title={A Novel Fractional Gradient-Based Learning Algorithm for Recurrent Neural Networks},

%   author={Khan, Shujaat and Ahmad, Jawwad and Naseem, Imran and Moinuddin, Muhammad},

%   journal={Circuits, Systems, and Signal Processing},

%   volume={37},

%   number={2},

%   pages={593--612},

%   year={2018},

%   publisher={Springer US}

% }

%% Description

% In this simulation I used a Golub et al(1999)'s Leukemia Cancer Database.

% The details of the dataset is available online at [1].  The leukemia db

% is a gene expression dataset contains 7128 genes, 2-classes (47-ALL &

% 25-AML), divided into two subsets training and test subsets. The training

% dataset contains 27-ALL, and 11-AML total 38 samples, and the test subset

% contains 20-ALL, and 14-AML total 34 samples.

%

% The genes are ranked using mRMR feature selection method [2] and the

% index of top 1000 genes is stored in 'feature_with_mRMr_d' vector.

% [1] http://www.stats.uwo.ca/faculty/aim/2015/9850/microarrays/FitMArray/chm/Golub.html

% [2] https://kr.mathworks.com/matlabcentral/fileexchange/14608-mrmr-feature-selection--using-mutual-information-computation-

%% Start

clc

clear all

close all

%% Load Golub et al. Leukemia Cancer DB

load Database/leukemia_dataset

feature_length=5;  % selecting top 5 most significant genes

train_data=train_data(feature_with_mRMr_d(1:feature_length),:)';

test_data=test_data(feature_with_mRMr_d(1:feature_length),:)';

%% Neural Network's parameters 

% (size/structure of the MLP)

N1=20;                  % Middle Layer Neurons

N2=1;                   % Output Layer Neurons 

N0=feature_length+1;    % Input Layer Neurons (feature length + bias)

% Training parameters

eta = 0.5; % Learning Rate 0-1   eg .01, .05, .005

epoch=10;  % Training iterations

⛄ 运行结果

⛄ 参考文献

[1]刘壮明. 前馈神经网络反向传播算法研究及在地面目标识别中的应用[J]. 声学所博硕士学位论文, 2006.

⛄ 完整代码

❤️部分理论引用网络文献，若有侵权联系博主删除

❤️ 关注我领取海量matlab电子书和数学建模资料