Prediction with Fuzzy Neural Network and Anfis

3 次查看（过去 30 天）

youngz 2016-7-15

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/295955-prediction-with-fuzzy-neural-network-and-anfis

编辑： youngz 2016-7-16

Hi everyone, I'am pretty new with Matlab and I'm trying to predict with a fuzzy neural net. Following some tutorials (finding on the web) I write this code:

M % complete matrix with a lot of row (>150000) and 39 column (the last one is the output of the nn, it could be 2 or 3)
TRAIN % 70% of M
CHECK % 15% of M
TEST % 15% of M
[Train_r Train_c] = size(TRAIN);
inp = TRAIN(:,1:(Train_c-1));
out = TRAIN(:,Train_c);
in_fis2 = genfis2(inp,out,0.5);
in_fis3 = genfis3(inp,out,'sugeno',3);
epoch_n = 20;
dispOpt = zeros(1,4);
out_fis2 = anfis(TRAIN,in_fis2,20,dispOpt);
out_fis3 = anfis(TRAIN,in_fis3,20,dispOpt);

Now I would like to predict these trainings on TEST and CHECK (like a classic ANN). After that, i would like to plot Confusion Matrix and Roc Curve. Do you have any suggestions?

Thanks

Edit: I see this example, however I would like to translate this tutorial in code!