Tried using starting values from an old kMeans run:
mus = [
[465440,190495,1000],
[526390,247188,1000],
[585512,333682,1000],
[649564,439416,1000],
[705580,591115,1000],
[827750,803499,1200]]
alphas = [1/K for _ in range(K)]
Sigmas = [np.cov([xs,ys,zs],rowvar=True) for _ in range(K)]
Sigma = [[0 for _ in range(d)] for _ in range(d)]
Sigma[0][0]= 1e9
Sigma[1][1]= 1e9
Sigma[2][2]= 1e4
Tried using starting values from an old kMeans run: mus = [ [465440,190495,1000], [526390,247188,1000], [585512,333682,1000], [649564,439416,1000], [705580,591115,1000], [827750,803499,1200]] alphas = [1/K for _ in range(K)]
Sigmas = [np.cov([xs,ys,zs],rowvar=True) for _ in range(K)]
Not successful