nf_test1.m

bp网络训练后的测试

%***Novelty Detection
%***Testing nf trained by 'nf_train'
%***Iput File: 1.prepared data file
%           or 2.Abaqus Data file '*.dat'
clear;clc;
%=====below can be changed by user====
  sample_test=500; %when data_from=1, this number is decided based on the size of prepared data file
  data_from=1;              
  data_file='DF_MS9B100.txt';  %%when data_from=1:get data from prepared data file
  %data_file='ksm1MScase8B.dat'; %%when data_from=2:get data from Abaqus Data file '*.dat'
  %load netwoek trained before by nf_train.m
  nf_trained=load('nftest_ksm1MS.mat');
%=====above can be changed by user====
novelty_filter=nf_trained.novelty_filter;
f=nf_trained.f;
alpha=nf_trained.alpha;
noise=nf_trained.noise;
inout_node=nf_trained.inout_node;
sample_train=nf_trained.sample_train;
lamda_train=zeros(1,sample_train);
lamda_test=zeros(1,sample_test);
for i=1:1:inout_node
   m=mean(f(i,:));
   y(i,:)=(f(i,:)-m)*alpha+m;
end
y1 = sim(novelty_filter,f);
for j=1:1:sample_train
   sum=0.0;
   for i=1:1:inout_node
      sum=sum+(y1(i,j)-y(i,j))^2.0;
   end
   lamda_train(j)=sqrt(sum);
end
threshold=mean(lamda_train)+4*std(lamda_train);
%***read test data from a prepared data file
if data_from==1 %1.get test data from a prepared data file
   fid=fopen(data_file,'r'); 
   ft=fscanf(fid,'%g',[inout_node,sample_test]);
   %read in order of column by column,
   %element number in a column is "inout_node"
   %total number of columns is "sample_test" 
   fclose(fid);
end  %data_from=1:get test data ft from prepared data file
if data_from==2
   %***read freq from abaqus.dat file and then add noise to form freq set
   number=-1;
   fin=fopen(data_file,'r');
   while number<=0
     line=fgetl(fin);
     matches=findstr(line,'CYCLES');
     number=length(matches);
   end
   line=fgetl(fin);
   line;
   for i=1:1:inout_node
     mode=fscanf(fin,'%d',1);
     x=fscanf(fin,'%g', 2);
     freq(i)=fscanf(fin,'%g', 1);
     z=fscanf(fin,'%g', 2);
   end 
   fclose(fin);
   freq=freq'; %change row to column
   %add noise to freq to form data set
   for i=1:1:inout_node
     ranf=sprandn([1:sample_test]); maxf=max(abs(ranf));
     ranf=ranf/maxf;
     for j=1:1:sample_test
       ft(i,j)=freq(i)+noise*ranf(j)*freq(i);
     end
   end
   %******ft is inout_node rows and samlp_test columns***********************************   
end %data_from=2:setup test data ft from Abaqus data file
%*****************************************   
for i=1:1:inout_node
   mt=mean(ft(i,:));
   yt(i,:)=(ft(i,:)-mt)*alpha+mt;
end
yt1=sim(novelty_filter,ft);
for j=1:1:sample_test
  sum=0.0;
    for i=1:1:inout_node
      sum=sum+(yt1(i,j)-yt(i,j))^2.0;
   end
   lamda_test(j)=sqrt(sum);
end

lamda=[lamda_train,lamda_test];
%ploting
figure(1);
%subplot(2,1,1);
plot(lamda,'b');
hold on;
plot([1,sample_train+sample_test],[threshold threshold],'-.k');
plot([sample_train sample_train],[0.8*min(lamda) 1.2*max(lamda)],'-.k');
xlabel('Training / testing data','fontname','times new roman','fontsize',16,'fontweight','bold');
ylabel('Novelty index','fontname','times new roman','fontsize',16,'fontweight','bold');
hold off;