Table 3 Dataset schema with detailed FSU specifications.

From: Enhanced sensing performance through the integration of denoising autoencoder and ensembling techniques

Column

Description

Type

Normal (\(H_0\))

Normal (\(H_1\))

NFS

YFS

1-10

Energy Readings

Float

N(k, 2k)

\(N(k(\eta +1),2k(\eta +1))\)

N(k, 2k)

\(N(k(\eta +1),2k(\eta +1))\)

11

PU State

Binary

\(H_0=0\)

\(H_1=1\)

\(H_0=0\) or \(H_1=1\)

\(H_0=0\) or \(H_1=1\)

  1. \(N(\mu ,\sigma ^2)\): Gaussian distribution; \(\eta\): SNR; k: scaling factor.
  2. Each row corresponds to one sensing interval; Labels are perfectly known; NFS: Always reports \(H_0\) energies; YFS: Always reports \(H_1\) energies.
Back to article page