GAN Peformance is very low for time series HAR data augmentation

15 Views Asked by At

I have 3D time series data X Y Z for human activity recognition. I am using GAN for data augmentation but the generated data is very different of original data and accuracy is also very low when i am using that data to test my base modal. I am using GAN implemented with fully connected layers(MLP).

0

There are 0 best solutions below