UNAGI.dynamic_graphs.distDistance

Functions

calculateKL(cluster1gaussian, cluster2gaussian)	calculate KL divergence of multivariate gaussian distributions between two clusters.
fitClusterGaussianRepresentation(data)	Fitting gaussian distributions for each hidden node
getClusterRepresentation(mus, sigmas)	MC strategy to sample gaussian data points.
getSimilarity(adatai, adataj, i, j)	calculate the differential gene similarity between two clusters.
mcSampling(mus, sigmas)	monte carlo sampling strategy to sample data points from a gaussian distribution in each cell, for example the hidden space is 10, then sample 100 data point from input cell.
normalizeDistance(distance)	Normalize the kl divergence distance and top differential gene distances.

Classes

GaussianRepThread(output, data, i)

The class to fit gaussian distributions for each hidden node.