Data and outputs for: How Many Degrees of Freedom Does a Neutrosophic Uncertainty Representation Have?
Annotatsiya
Stored network outputs and trained weights behind every number in the article "How Many Degrees of Freedom Does a Neutrosophic Uncertainty Representation Have? Theory, and Tests on Hyperspectral Convolutional Networks and Preference Data": test-pixel outputs of the 70 prespecified hyperspectral runs (Pavia University, Salinas), the transductive-preprocessing runs, the revision runs (retrained replicas with weights, NS-U with head penalty, alternative evidential objectives, covariate shift) and the CIFAR-10H positive control. The code, the tables computed from these files and a map from every result to its script are in the GitHub repository. Third-party data is not included. See README_DATA.md.
Hali tarjima qilinmagan