《MATLAB 神经网络43个案例分析》：第43章 神经网络高效编程技巧——基于MATLAB R2012b新版本特性的探讨
1. 前言
《MATLAB 神经网络43个案例分析》是MATLAB技术论坛（www.matlabsky.com）策划，由王小川老师主导，2013年北京航空航天大学出版社出版的关于MATLAB为工具的一本MATLAB实例教学书籍，是在《MATLAB神经网络30个案例分析》的基础上修改、补充而成的，秉承着“理论讲解—案例分析—应用扩展”这一特色，帮助读者更加直观、生动地学习神经网络。
《MATLAB神经网络43个案例分析》共有43章，内容涵盖常见的神经网络（BP、RBF、SOM、Hopfield、Elman、LVQ、Kohonen、GRNN、NARX等）以及相关智能算法（SVM、决策树、随机森林、极限学习机等）。同时，部分章节也涉及了常见的优化算法（遗传算法、蚁群算法等）与神经网络的结合问题。此外，《MATLAB神经网络43个案例分析》还介绍了MATLAB R2012b中神经网络工具箱的新增功能与特性，如神经网络并行计算、定制神经网络、神经网络高效编程等。
近年来随着人工智能研究的兴起，神经网络这个相关方向也迎来了又一阵研究热潮，由于其在信号处理领域中的不俗表现，神经网络方法也在不断深入应用到语音和图像方向的各种应用当中，本文结合书中案例，对其进行仿真实现，也算是进行一次重新学习，希望可以温故知新，加强并提升自己对神经网络这一方法在各领域中应用的理解与实践。自己正好在多抓鱼上入手了这本书，下面开始进行仿真示例，主要以介绍各章节中源码应用示例为主，本文主要基于MATLAB2015b(32位)平台仿真实现，这是本书第四十三章神经网络高效编程技巧实例，话不多说，开始！
2. MATLAB 仿真示例
打开MATLAB，点击“主页”，点击“打开”，找到示例文件
选中chapter43.m，点击“打开”
chapter43.m源码如下：
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%功能：神经网络高效编程技巧-基于MATLAB R2012b新版本特性的探讨
%环境：Win7，Matlab2015b
%Modi: C.S
%时间：2022-06-21
%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%
%% Matlab神经网络43个案例分析
% 神经网络高效编程技巧-基于MATLAB R2012b新版本特性的探讨
% by 王小川(@王小川_matlab)
% http://www.matlabsky.com
% Email:sina363@163.com
% http://weibo.com/hgsz2003
% 注意代码为演示脚本，请根据备注分块运行
close all;
clear all
clc
tic
%% 典型的前向神经网络
[x,t]=house_dataset;
net1=feedforwardnet(10);
net2=train(net1,x,t);
y=net2(x);
%% 设置reduction
net=train(net1,x,t,'reduction',10);
y=net(x,'reduction',10);
%% 神经网络并行运算
% CPU并行
% matlabpool open
delete(gcp('nocreate'))
numWorkers = parpool(2)
[x,t]=house_dataset;
net=feedforwardnet(10);
net=train(net,x,t,'useparallel','yes')
y=sim(net,x,'useparallel','yes')
% GPU并行
net=train(net,x,t,'useGPU','yes')
y=sim(net,x,'useGPU','yes')
%% Elliot S函数的使用
n=-10:0.01:10;
a1=elliotsig(n);
a2=tansig(n);
h=plot(n,a1,n,a2);
legend(h,'ELLIOTSIG','TANSIG','location','NorthWest')
n = rand(1000,1000);
tic, for i=1:100, a = elliotsig(n); end, elliotsigTime = toc
tic, for i=1:100, a = tansig(n); end, tansigTime = toc
speedup = tansigTime / elliotsigTime
%% 神经网络负载均衡
[x,t] = house_dataset;
Xc = Composite;
Tc = Composite;
Xc{
1} = x(:, 1:100); % First 100 samples of x
Tc{
1} = t(:, 1:100); % First 100 samples of t
Xc{
2} = x(:, 101:506); % Rest samples of x
Tc{
2} = t(:, 101:506); % Rest samples of t
%% 代码组织更新
help nnprocess
help nnweight
help nntransfer
help nnnetinput
help nnperformance
help nndistance
toc
添加完毕，点击“运行”，开始仿真，输出仿真结果如下：
Parallel pool using the 'local' profile is shutting down.
Starting parallel pool (parpool) using the 'local' profile ...
connected to 2 workers.
numWorkers =
Pool - 属性:
Connected: true
NumWorkers: 2
Cluster: local
AttachedFiles: {
}
AutoAddClientPath: true
IdleTimeout: 30 minutes (30 minutes remaining)
SpmdEnabled: true
net =
Neural Network
name: 'Feed-Forward Neural Network'
userdata: (your custom info)
dimensions:
numInputs: 1
numLayers: 2
numOutputs: 1
numInputDelays: 0
numLayerDelays: 0
numFeedbackDelays: 0
numWeightElements: 151
sampleTime: 1
connections:
biasConnect: [1; 1]
inputConnect: [1; 0]
layerConnect: [0 0; 1 0]
outputConnect: [0 1]
subobjects:
input: Equivalent to inputs{
1}
output: Equivalent to outputs{
2}
inputs: {
1x1 cell array of 1 input}
layers: {
2x1 cell array of 2 layers}
outputs: {
1x2 cell array of 1 output}
biases: {
2x1 cell array of 2 biases}
inputWeights: {
2x1 cell array of 1 weight}
layerWeights: {
2x2 cell array of 1 weight}
functions:
adaptFcn: 'adaptwb'
adaptParam: (none)
derivFcn: 'defaultderiv'
divideFcn: 'dividerand'
divideParam: .trainRatio, .valRatio, .testRatio
divideMode: 'sample'
initFcn: 'initlay'
performFcn: 'mse'
performParam: .regularization, .normalization
plotFcns: {
'plotperform', plottrainstate, ploterrhist,
plotregression}
plotParams: {
1x4 cell array of 4 params}
trainFcn: 'trainlm'
trainParam: .showWindow, .showCommandLine, .show, .epochs,
.time, .goal, .min_grad, .max_fail, .mu, .mu_dec,
.mu_inc, .mu_max
weight and bias values:
IW: {
2x1 cell} containing 1 input weight matrix
LW: {
2x2 cell} containing 1 layer weight matrix
b: {
2x1 cell} containing 2 bias vectors
methods:
adapt: Learn while in continuous use
configure: Configure inputs & outputs
gensim: Generate Simulink model
init: Initialize weights & biases
perform: Calculate performance
sim: Evaluate network outputs given inputs
train: Train network with examples
view: View diagram
unconfigure: Unconfigure inputs & outputs
y =
1 至 11 列
26.5649 23.8963 35.3167 34.0331 33.2050 26.6177 19.4666 17.6496 16.9333 17.5074 17.8289
12 至 22 列
18.4232 20.2884 19.2130 18.1789 19.2407 21.1541 17.4378 18.4759 18.1597 15.8114 17.2226
23 至 33 列
17.2158 16.0054 16.4798 15.9178 16.3531 16.5772 18.5624 19.9107 15.7658 17.0476 16.2111
34 至 44 列
16.0783 16.2199 20.9717 21.3573 22.4374 24.5116 31.6593 37.3553 30.6522 23.5682 23.8258
45 至 55 列
22.8457 21.6117 22.9742 18.7009 17.5504 19.2179 20.7874 22.4802 26.8306 21.2091 23.8159
56 至 66 列
37.3504 25.1582 31.1299 23.4489 21.9053 19.0342 17.4677 23.4439 25.3006 30.9247 28.1945
67 至 77 列
22.5599 22.0387 19.7366 21.3109 26.9079 22.4068 24.9839 25.2604 29.6492 22.1388 18.9827
78 至 88 列
21.7213 19.7240 21.6383 28.7301 23.4959 25.2390 22.9466 24.8965 27.2614 24.2773 24.2914
89 至 99 列
28.2512 31.0355 25.2934 24.2401 24.7488 27.3502 20.3813 26.9904 23.1447 40.5675 44.5740
100 至 110 列
34.1028 22.8109 23.9780 21.2773 19.9033 20.0861 18.9174 19.4450 20.6953 20.5032 19.9361
111 至 121 列
23.8382 21.7507 18.5984 18.5463 20.1065 17.7629 20.5600 18.9054 18.9263 20.2244 21.2653
122 至 132 列
21.9477 19.2307 16.7994 19.0573 21.2896 16.4173 16.8229 19.1419 16.0063 20.3104 19.8476
133 至 143 列
20.7054 16.9056 15.5118 17.6302 17.1318 19.5664 15.9773 17.4037 15.7664 13.6104 14.2766
144 至 154 列
16.8868 15.0215 7.7864 15.7275 14.3187 13.1964 16.7919 20.9003 18.7391 14.3712 17.3107
155 至 165 列
17.2587 17.5046 13.5260 38.6815 28.2783 25.5819 29.7610 49.1533 47.4279 51.0366 22.1065
166 至 176 列
23.6777 50.0421 20.4957 23.4337 23.8454 18.4839 20.5531 23.6133 25.1760 24.7953 29.9514
177 至 187 列
25.4153 24.4943 28.8963 37.5006 41.4320 29.9014 35.9078 29.6126 23.4604 27.7544 46.9291
188 至 198 列
32.0717 29.8403 34.1304 30.3110 27.6823 33.9577 32.6796 29.8632 48.0170 37.3538 31.3165
199 至 209 列
35.0443 34.7285 35.7639 24.4859 42.1004 48.5204 51.6130 22.4677 21.9444 18.1892 21.9470
210 至 220 列
17.7342 20.7005 18.0920 20.6546 24.4089 27.7455 22.6432 24.4801 25.0930 18.9495 28.9056
221 至 231 列
26.7565 21.5714 26.1969 24.7401 43.4779 47.2852 40.4343 29.2835 45.7140 32.3809 22.4439
232 至 242 列
32.2503 44.9441 43.2386 26.8292 23.5077 25.0582 32.1496 25.7791 25.4367 25.2751 20.2267
243 至 253 列
21.5403 24.3852 17.0162 16.6969 21.7378 19.2453 22.6710 26.1453 25.8987 26.7321 26.9997
254 至 264 列
36.8934 22.9069 19.3476 41.6412 54.1393 36.7100 32.0374 34.7428 39.4294 48.8169 34.8423
265 至 275 列
35.5150 23.1172 29.9240 50.3426 47.1693 18.7587 22.1208 25.4768 26.0070 37.2942 32.9633
276 至 286 列
32.1191 34.1121 33.0264 28.0307 37.2538 46.2947 37.4417 43.9732 53.1790 32.5625 25.4412
287 至 297 列
20.5018 23.1884 23.1355 25.3114 35.6611 38.2709 33.5755 24.2597 22.5439 30.9491 28.5340
298 至 308 列
19.2980 24.7290 32.5504 27.3564 26.2345 27.5806 33.5211 36.9534 30.7142 36.4146 30.9553
309 至 319 列
24.2694 18.6600 18.0868 22.5957 18.1427 19.9788 20.6990 17.0862 17.2648 17.7313 20.5753
320 至 330 列
19.4429 23.5494 23.1678 21.3442 17.7827 24.5917 24.7210 23.4610 20.6296 22.4372 24.7429
331 至 341 列
20.9764 19.2289 21.0957 23.3897 22.7596 20.7233 20.4455 19.6999 22.2372 21.5693 20.8254
342 至 352 列
33.8562 19.6238 25.9744 31.4417 18.3665 17.4625 29.1199 28.0675 29.4162 24.9090 25.7243
353 至 363 列
19.7282 31.0909 18.9530 24.5006 19.2800 18.3343 15.2951 22.6964 23.9613 21.6809 22.9029
364 至 374 列
20.4064 23.7425 34.3384 24.2606 33.7902 50.6635 57.1621 57.2010 36.1846 46.4003 9.5275
375 至 385 列
5.6477 20.9966 15.0029 15.6707 11.8745 11.9566 10.6016 12.5754 9.9299 9.9765 8.4788
386 至 396 列
7.5003 3.5342 9.7833 5.1151 11.7370 16.1919 19.0900 7.0987 20.7425 16.6463 16.9238
397 至 407 列
16.7475 14.1719 4.5413 17.2885 6.7930 12.2707 13.4259 8.8577 11.2502 3.1350 12.0974
408 至 418 列
28.4709 19.2913 19.4778 18.4398 18.9188 15.3099 17.9595 3.5736 8.9826 9.7391 8.9614
419 至 429 列
5.4574 11.0193 17.9750 19.4023 22.6472 13.3676 16.3755 10.2455 14.7707 10.0486 12.2779
430 至 440 列
10.8847 18.6243 17.2836 21.6089 16.2555 14.8827 10.0539 12.3158 8.2350 7.8187 12.0014
441 至 451 列
10.8312 13.7364 16.6852 13.6024 9.7569 9.5413 16.7027 16.0780 12.0619 13.1249 13.2298
452 至 462 列
14.6964 17.0095 15.1535 11.4741 13.7664 12.2353 12.9877 17.3928 19.7702 17.3746 20.4944
463 至 473 列
21.1293 22.6934 23.5661 23.5699 13.5559 15.4246 14.6432 22.0109 16.3024 23.3346 20.9164
474 至 484 列
26.2760 18.4605 15.4811 15.5182 10.0277 15.1480 22.2057 23.2604 26.9101 26.8724 24.2792
485 至 495 列
23.7287 23.7627 14.8351 25.5489 13.2138 5.5286 -0.9328 12.9850 16.7500 21.8281 24.9077
496 至 506 列
23.7792 19.6660 18.9583 21.3667 19.2257 19.5222 21.4640 18.9323 25.4267 23.4556 18.7017
NOTICE: Jacobian training not supported on GPU. Training function set to TRAINSCG.
net =
Neural Network
name: 'Feed-Forward Neural Network'
userdata: (your custom info)
dimensions:
numInputs: 1
numLayers: 2
numOutputs: 1
numInputDelays: 0
numLayerDelays: 0
numFeedbackDelays: 0
numWeightElements: 151
sampleTime: 1
connections:
biasConnect: [1; 1]
inputConnect: [1; 0]
layerConnect: [0 0; 1 0]
outputConnect: [0 1]
subobjects:
input: Equivalent to inputs{
1}
output: Equivalent to outputs{
2}
inputs: {
1x1 cell array of 1 input}
layers: {
2x1 cell array of 2 layers}
outputs: {
1x2 cell array of 1 output}
biases: {
2x1 cell array of 2 biases}
inputWeights: {
2x1 cell array of 1 weight}
layerWeights: {
2x2 cell array of 1 weight}
functions:
adaptFcn: 'adaptwb'
adaptParam: (none)
derivFcn: 'defaultderiv'
divideFcn: 'dividerand'
divideParam: .trainRatio, .valRatio, .testRatio
divideMode: 'sample'
initFcn: 'initlay'
performFcn: 'mse'
performParam: .regularization, .normalization
plotFcns: {
'plotperform', plottrainstate, ploterrhist,
plotregression}
plotParams: {
1x4 cell array of 4 params}
trainFcn: 'trainscg'
trainParam: .showWindow, .showCommandLine, .show, .epochs,
.time, .goal, .min_grad, .max_fail, .sigma,
.lambda
weight and bias values:
IW: {
2x1 cell} containing 1 input weight matrix
LW: {
2x2 cell} containing 1 layer weight matrix
b: {
2x1 cell} containing 2 bias vectors
methods:
adapt: Learn while in continuous use
configure: Configure inputs & outputs
gensim: Generate Simulink model
init: Initialize weights & biases
perform: Calculate performance
sim: Evaluate network outputs given inputs
train: Train network with examples
view: View diagram
unconfigure: Unconfigure inputs & outputs
y =
1 至 11 列
26.1052 23.9302 35.3771 33.8291 33.0630 26.3830 19.0359 17.5569 16.8636 17.2882 17.7174
12 至 22 列
18.2294 19.6969 18.9601 18.1081 18.8878 20.4445 17.3256 17.8212 17.9337 15.7446 17.1369
23 至 33 列
17.1539 15.9448 16.4072 15.7874 16.2701 16.4893 18.5389 19.8947 15.6949 16.9973 16.0209
34 至 44 列
15.9942 16.1378 20.6872 20.9464 21.7688 23.8159 29.7893 35.4520 30.0925 22.9551 23.2047
45 至 55 列
22.3536 21.0165 22.4210 18.4536 17.3079 18.6898 20.1176 22.0121 26.0986 20.4135 22.7862
56 至 66 列
35.7018 23.5710 29.4466 22.6233 21.1450 18.4573 17.2552 23.0774 24.6646 30.6855 26.2513
67 至 77 列
21.0169 21.1902 18.9905 20.5382 26.1687 21.6999 24.1548 24.5064 28.6841 21.8397 18.9234
78 至 88 列
21.3944 19.5442 21.1502 27.8482 23.1162 24.2972 22.1583 24.4150 26.9827 23.7395 23.8799
89 至 99 列
28.5188 30.9379 25.0518 24.1241 24.1114 26.0978 20.1648 26.7065 22.8512 40.9028 44.2740
100 至 110 列
34.0684 22.8012 23.9517 21.2881 19.8664 20.0481 18.9029 19.4613 20.6748 20.4826 19.9206
111 至 121 列
23.6942 21.7098 18.6340 18.5892 20.0501 17.7523 20.4816 18.8208 18.8787 20.1178 21.0919
122 至 132 列
21.9772 19.2721 16.8112 19.0977 21.3317 16.4186 17.0111 19.3549 16.2832 20.4699 20.0176
133 至 143 列
20.8380 17.1623 15.7846 17.9559 17.4110 19.7630 16.2902 17.6726 16.3440 14.2640 14.0424
144 至 154 列
17.3563 15.7702 8.2714 15.2524 15.1402 13.9131 16.7177 20.2764 17.9912 13.7273 16.7480
155 至 165 列
16.6946 16.9625 13.1041 38.2405 27.8189 24.7960 29.1962 48.5739 47.1319 50.8078 21.7652
166 至 176 列
23.4894 49.6868 20.1898 23.2567 23.6505 18.3392 20.3294 23.7072 25.2985 24.6221 29.4758
177 至 187 列
24.9921 24.4312 28.9922 37.4922 41.8402 29.7498 36.3731 29.9761 23.3787 27.5658 46.8859
188 至 198 列
30.8714 28.3625 32.8117 28.7918 26.3096 32.4524 31.1045 28.4310 46.4336 35.7800 29.8718
199 至 209 列
33.5372 32.8610 33.9137 22.8331 40.0832 46.7982 49.8822 21.7295 21.5281 17.9203 21.9977
210 至 220 列
17.8236 20.8489 18.2242 20.6413 23.7560 27.8925 22.0683 23.9388 25.2703 19.2414 28.8964
221 至 231 列
27.0582 21.9844 26.4298 24.8221 43.9769 47.8780 40.9457 29.5095 45.1444 31.6332 22.2249
232 至 242 列
32.5266 45.3645 43.6243 26.8567 23.2040 25.2528 32.3439 24.7259 24.6634 24.6140 19.6113
243 至 253 列
20.8490 23.2388 16.6810 16.2319 21.0220 19.0348 22.0843 25.3862 24.9961 25.7506 26.3620
254 至 264 列
36.5440 21.4507 17.8251 39.9753 54.2588 37.1643 32.4035 34.9698 39.7393 49.0541 35.2641
265 至 275 列
35.8699 23.0767 30.1698 50.3990 46.9021 18.6542 21.3872 24.4769 25.4980 37.2529 31.8644
276 至 286 列
31.1094 33.5341 31.9123 26.8866 37.1073 46.1695 37.1640 43.8987 51.6298 30.7881 24.2471
287 至 297 列
19.1356 22.0448 22.1756 24.0929 33.8793 36.5313 31.7265 23.6091 22.0843 30.5687 28.4142
298 至 308 列
18.8895 23.1113 30.7386 26.0831 25.3413 26.4507 32.2505 36.1615 30.0673 36.0605 30.4683
309 至 319 列
24.3732 18.5362 17.4719 22.0675 18.1226 19.9667 20.7546 16.9316 17.1622 17.5083 20.3804
320 至 330 列
19.0718 23.1399 22.7882 20.8261 17.5458 24.0109 23.9426 22.7860 20.0667 21.6882 23.8885
331 至 341 列
20.2886 18.2118 20.0401 22.8820 22.2405 20.0701 19.8085 19.1981 21.5486 20.9327 20.3238
342 至 352 列
33.1751 19.2608 24.8264 29.7636 17.9061 17.0023 27.4173 26.4054 28.2609 23.8518 24.5448
353 至 363 列
18.5078 29.8888 17.7396 23.1480 18.1294 16.9033 13.7941 21.0323 22.1927 20.2551 21.7974
364 至 374 列
18.9514 21.6398 33.1362 23.2493 32.6777 49.5711 54.8847 54.9243 35.6095 44.3276 12.2624
375 至 385 列
8.7500 20.5041 15.3502 16.2838 12.8156 12.9389 11.0646 13.2939 11.1418 11.3080 9.3977
386 至 396 列
9.4044 5.5720 11.4089 7.0473 12.6022 16.2866 18.3116 8.5392 20.2570 16.5592 16.8953
397 至 407 列
16.8658 14.8100 6.4503 17.4849 8.1966 12.8346 13.7658 9.5697 11.7624 4.4385 14.0421
408 至 418 列
28.0355 20.5128 18.8668 16.7771 17.7931 14.9449 17.9058 3.7976 8.3035 8.8910 8.4500
419 至 429 列
4.7895 9.8625 17.1620 18.5163 21.9529 12.3248 14.9401 9.3703 13.2836 8.7765 11.1714
430 至 440 列
10.0530 17.4321 16.2562 19.9220 14.7849 13.5412 9.1382 11.0436 7.4455 7.2669 12.2955
441 至 451 列
10.9708 13.4480 16.3358 13.3781 9.3123 8.6123 15.8215 15.4831 12.0833 12.6201 11.9444
452 至 462 列
14.1960 16.3897 14.5162 10.3416 12.4374 11.0103 11.5851 16.0108 18.6785 16.0854 19.4676
463 至 473 列
19.8734 21.4142 22.0120 21.6229 12.3140 16.0234 14.2611 20.8949 16.3751 23.6808 20.2390
474 至 484 列
24.7399 19.2356 16.1301 16.0267 11.3811 15.7718 21.8281 22.2273 25.8419 25.7634 22.5569
485 至 495 列
21.9501 22.0320 14.5170 23.9251 15.4170 8.1912 2.2122 15.2351 18.8811 21.5379 24.5755
496 至 506 列
23.7203 19.7834 18.8922 21.2567 19.1922 19.5051 21.3002 18.7973 25.5533 23.5363 18.6050
elliotsigTime =
0.4284
tansigTime =
0.8825
speedup =
2.0603
Neural Network Toolbox Processing Functions.
General Data Preprocessing
fixunknowns - Processes matrix rows with unknown values.
mapminmax - Map matrix row minimum and maximum values to [-1 1].
mapstd - Map matrix row means and deviations to standard values.
processpca - Processes rows of matrix with principal component analysis.
removeconstantrows - Remove matrix rows with constant values.
removerows - Remove matrix rows with specified indices.
Data Preprocessing for Specific Algorithms
lvqoutputs - Define settings for LVQ outputs, without changing values.
Main nnet function list.
Neural Network Toolbox Weight Functions.
Weight functions
convwf - Convolution weight function.
dotprod - Dot product weight function.
negdist - Negative distance weight function.
normprod - Normalized dot product weight function.
scalprod - Scalar product weight function.
Distance functions can be used as weight functions
boxdist - Box distance function.
dist - Euclidean distance weight function.
linkdist - Link distance function.
mandist - Manhattan distance function.
Main nnet function list.
Neural Network Toolbox Transfer Functions.
compet - Competitive transfer function.
elliotsig - Elliot sigmoid transfer function.
hardlim - Positive hard limit transfer function.
hardlims - Symmetric hard limit transfer function.
logsig - Logarithmic sigmoid transfer function.
netinv - Inverse transfer function.
poslin - Positive linear transfer function.
purelin - Linear transfer function.
radbas - Radial basis transfer function.
radbasn - Radial basis normalized transfer function.
satlin - Positive saturating linear transfer function.
satlins - Symmetric saturating linear transfer function.
softmax - Soft max transfer function.
tansig - Symmetric sigmoid transfer function.
tribas - Triangular basis transfer function.
Main nnet function list.
Neural Network Toolbox Net Input Functions.
netprod - Product net input function.
netsum - Sum net input function.
Main nnet function list.
Neural Network Toolbox Performance Functions.
mae - Mean absolute error performance function.
mse - Mean squared error performance function.
sae - Sum absolute error performance function.
sse - Sum squared error performance function.
crossentropy - Cross-entropy performance.
msesparse - Mean squared error performance function with L2 weight and sparsity regularizers.
Main nnet function list.
Neural Network Toolbox Distance Functions.
boxdist - Box distance function.
dist - Euclidean distance weight function.
linkdist - Link distance function.
mandist - Manhattan distance function.
Main nnet function list.
时间已过 3.748946 秒。
依次点击Plots框下的Performance，Training State,Error Histogram,Regression可得到以下图示：
3. 小结
并行计算相关函数与数据调用在MATLAB2015版本之后都有升级，所以需要对源码进行调整再进行仿真，本文示例均根据机器情况与MATLAB版本进行适配。对本章内容感兴趣或者想充分学习了解的，建议去研习书中第四十三章节的内容。后期会对其中一些知识点在自己理解的基础上进行补充，欢迎大家一起学习交流。