from scipy import linalg as spl from scipy import stats as sps from matplotlib import pyplot as plt import GPdc import PES nt=12 d=1 lb = sp.array([-1.]*d) ub = sp.array([1.]*d) [X,Y,S,D] = ESutils.gen_dataset(nt,d,lb,ub,GPdc.SQUEXP,sp.array([1.5,0.15])) G = PES.makeG(X,Y,S,D,GPdc.SQUEXP,sp.array([0.,-1.]),sp.array([1.,1.]),12) Z=PES.drawmins(G,8,sp.array([-1.]),sp.array([1.]),SUPPORT=400,SLICELCB_PARA=1.) Ga = GPdc.GPcore(*PES.addmins(G,X,Y,S,D,Z[0,:])+[G.kf]) np=100 sup = sp.linspace(-1,1,np) Dp = [[sp.NaN]]*np Xp = sp.vstack([sp.array([i]) for i in sup]) [m,V] = G.infer_diag_post(Xp,Dp) [mp,Vp] = Ga.infer_diag_post(Xp,Dp) f,a = plt.subplots(2) s = sp.sqrt(V[0,:]) a[0].fill_between(sup,sp.array(m[0,:]-2.*s).flatten(),sp.array(m[0,:]+2.*s).flatten(),facecolor='lightblue',edgecolor='lightblue') a[0].plot(sup,m[0,:].flatten()) a[0].plot(sp.array(X[:,0]).flatten(),Y,'g.')
