神经网络仿真模块的设计编写
- <html xmlns="http://www.w3.org/DTD1999/xhtml">
- <html>
- <head>
- <meta http-equiv="content-Iype" content="text/html;
- charset=GB2312">
- <a href="http://www.cmpbook.com/images">
- <title>神经网络系统</title>
- <script dd="stylesheet" _blank=br object createpopup buttom="设计神经网络"></a>
- load netL7.mat %加载net7.mat文件
- 然后再输入命令:
- net
- 以输出的预设值为:
- net=
- Neural Network object:
- 表决以应用的函数预设值列序为:
- architecture
- numlnputs:1
- biasConnect:[1;1]
- inputconnect:[1;0]
- layerConnect:[0 0;1 0]
- outputConnect:[0 1]
- numOutputs:1 (read-only)
- numlnputDelays:0 (read-only)
- numLayerDelays:0 (read-only)
- 加载生成神经网络的函数预设值所得运行程序依次入列序
- suboject stuctures:
- inputs:{1×1 cell} of inputs
- layers:{2×1 cell} of layers
- outputs:{1×2 cell} containing 1 output
- biases:{2×1 cell} containing 2 biases
- inputWeight:{2×1 cell} containing 1 input weight
- layerWeights:{2×2 cell} containing 1 layer weight
- functions:
- adaptFen:'trains'
- divideFcn:(none)
- gradientFcn:'gdefaults'
- initFcn:'initlay'
- performFcn:'mse'
- plotFcns:{'plotperform','plottrainstate','plotregression'}
- trainFcn:'trainbfg'
- parameters:
- 生成列序神经网络的加载预设值产生的结果,形成列序预设值产生加载在window中
- adaptParam:passes
- divideParam:(none)
- gradienParam:(none)
- initParam:(none)
- performParam:(none)
- tainParam:.show,.showWindow,.showCommandLine,.epochs,
- .time,.goal,.max_fail,.min_beta,
- .searchFcn,.scale_tol,.alpha,.beta,
- .delta,.gama,.low_lim,.up_lim,
- .maxstep,.minstep,.bmax,.batch_frag
- weight and bias values:
- IW:{2×1 cell} containing 1 input weight matrix
- LW:{2×2 cell} containing 1 layer weight matrix
- b:{2×1 cell} containing 2 bias vectors
- other:
- name:"
- userdata:(user information)
- 在命令窗口加载设计好的神经网络后,以gensim函数生成神经网络仿真模块,在命令窗口中输入:
- gensim(net,-1) 注意: 仿真模块在设计过程中也是有一定的设想,比如要建设一座高楼大夏建筑工程好必要启用一个牢固的地层地基, 把关好地质基础,才会建成一座很好的漂亮让人耳目一新的高楼。
本文转自huangyouliang10 51CTO博客,原文链接:http://blog.51cto.com/1572091hyl10/575117