Answers are grounded in this paper and cite their sources.
The four-component (T, I, N, F) neutrosophic representation distinguishes pure indeterminacy I from neutrality N, in addition to truth T and falsehood F. Existing neutrosophic-graph research mainly assigns uncertain values to vertices or edges and studies the resulting graph structure. This paper addresses a different problem: a fixed weighted network is given, while a (T, I, N, F) assessment is the signal carried by each vertex. A benefit-coordinate embedding converts the semantic order into the standard product order and permits graph-Laplacian methods without merging neutrality and indeterminacy. For continuous-time heat diffusion, we prove preservation of the TINF cube, semantic order, component means, complement symmetry, and a maximum principle. A weighted Dirichlet energy is shown to decay monotonically, with exponential convergence governed by the algebraic connectivity. A stable discrete-time scheme and a soft-anchor graph filter are also established. For partially observed signals, the TINF harmonic extension is characterized simultaneously as the unique Dirichlet-energy minimizer and as a convex combination of boundary assessments. This yields complement equivariance, a semantic maximum principle, nonexpansive dependence on boundary data, and an explicit topology-perturbation bound. Complement-paired TINF prototypes are then introduced for four-state semi-supervised inference. They form a regular simplex in benefit coordinates, and nearest-prototype TINF classification is proved exactly equivalent to classical multiclass harmonic label propagation when anchors use the prototypes. In contrast, a symmetric three-component collapse makes neutrality and pure indeterminacy non-identifiable. An exact cycle example and Monte Carlo stochastic-block experiments confirm the equivalence numerically and demonstrate the information loss caused by collapse. The framework provides a mathematically auditable route for propagating four-component information over social, sensor, and knowledge networks.
No takes yet. Share an insight, caveat, or question.
Aqib et al. (2026) studied this question.
Synapse has enriched 8 closely related papers on similar clinical questions. Consider them for comparative context: