I want highest possible accuracy in my Training network

5 views (last 30 days)
Attached file is a dataset 297 x 13 measurements and 1 output as( 0, 1 and 2)
i want to train 75 % of datatraining with 25% test .
when i try my code i could get only 60 % accuracy.
traininput=traininput';
traintarget=traintarget';
testinput=testinput';
net=feedforwardnet(10); % have tried multiple layers
net=train(net.traininput,traintarget);
y=net(testinput);
output=y';
i have tried many things since a week but unable to get highest accuracy.
can anybody help please?

Answers (1)

Shashank Gupta
Shashank Gupta on 12 Apr 2021
Hi,
I haven't tried training your data, but I can give you suggestion on how to improve the accuracy, It generally takes time to find the most optimise network. I would suggest to first train an SVM model and check the accuracy. If the SVM doesn't work as per your expectation then go for more complex neural networks. May be look at this link. There is no hard and fast rule for finding out the optimised network, you just have to visualise the data and act accordingly.
I hope this helps or atleast give an heastart.
Cheers.

Categories

Find more on Sequence and Numeric Feature Data Workflows in Help Center and File Exchange

Products

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!