Advertisement
Not a member of Pastebin yet?
Sign Up,
it unlocks many cool features!
- load winequalityred.txt
- datainnet=[winequalityred(:,1:11)]';
- dataoutnet=[zeros(11,1599);]
- for i=1:1599
- skupina=winequalityred(i,12);
- dataoutnet(skupina,i)=1;
- end
- pocet_neuronov=220; % definujte pocet neuronov v skrytej vstve
- net = patternnet(pocet_neuronov);
- net.divideFcn='dividerand';
- net.divideParam.trainRatio=0.6;
- net.divideParam.valRatio=0.0;
- net.divideParam.testRatio=0.4;
- % nastavenie parametrov trenovania
- % definujte parametre trenovania siete
- net.trainParam.goal = 1e-18;
- net.trainParam.min_grad = 1e-18;
- % ukoncovacia podmienka na chybu.
- net.trainParam.show = 10; % frekvencia zobrazovania chyby
- net.trainParam.epochs = 6000;
- net = train(net,datainnet,dataoutnet);
- view(net)
- % simulacia vystupu NS pre trenovacie data
- % testovanie NS
- outnetsim = sim(net,datainnet);
- % chyba NS a dat
- err=(outnetsim-dataoutnet);
- % percento neuspesne klasifikovanych bodov
- % pouzit funkciu -> confusion
- disp(confusion(dataoutnet,outnetsim))
- %kontingenčná matica
- figure
- % pouzit funkciu -> plotconfusion
- plotconfusion(dataoutnet,outnetsim)
Advertisement
Add Comment
Please, Sign In to add comment
Advertisement