Gaussian Mixture Regression Github. Hueber T Girin L Alameda-Pineda X Bailly G. 2015 Speaker-Adaptive Acoustic-Articulatory Inversion using Cascaded Gaussian Mixture Regression in IEEEACM Transactions on Audio Speech and.
Gaussian processing GP is quite a useful technique that enables a non-parametric Bayesian approach to modeling. π For each observation i and Gaussian j set the weight in the E step. μ Σ p Δ i j.
This code is for paper titled An efficient sparse Bayesian learning algorithm Based on Gaussian-scale mixtures.
For prediction it uses a weighted conditional Gaussian distribution based on the trained GMM model. π For each observation i and Gaussian j set the weight in the E step. μ Σ p Δ i j. As a result the obtained affinity matrix is better at characterizing the structure of data in real applications.
