Y. Shrivastava

Cross-load robustness evaluation of deep and handcrafted learning frameworks for bearing fault diagnosis

This paper introduces a comprehensive leakage-free methodology for bearing fault diagnosis by three of the most representative feature learning approaches: Raw Signal Convolutional Neural Network (Raw CNN), Fast Fourier Transform (FFT) based hybrid MLP-CNN, and handcrafted statistical descriptor based Multilayer Percep...