K-Kernel Fit to the Real SPARC Galaxy NGC3198: Evidence Against Dark Matter

This paper presents a direct fit of the non-Markovian K-Kernel model to real observational data from the SPARC (Spitzer Photometry and Accurate Rotation Curves) database. Specifically, we analyze the galaxy NGC3198, a classic test case for Dark Matter. Using Markov Chain Monte Carlo (MCMC) with the emcee sampler, we fit three parameters: the mass-to-light ratio (M/L), the memory decay rate (lambda_k), and the memory amplitude (k_amp). The K-Kernel model reproduces the flat rotation curve of NGC3198 with high accuracy. The M/L ratio converges to a physically realistic value (1.026), confirming that no Dark Matter halo is required. This provides direct observational evidence that the non-Markovian memory of spacetime is the true cause of galactic rotation dynamics.