LogLike = LogLike1 + LogLike2 - CoordTransTerm print "calculate likelihood: {0} s".format(time.time()-tstart) , LogLike #print 'EFAC:', np.abs(p0) #print 'ECORR:', np.abs(p2) #print 'RNAMP: %s, RNIDX: %s' % (np.exp(p3[0]), p3[1]) print 'plist:', plist return LogLike LogLike = lambda x:loglikelihood(x) * -1. from testmcmc import SliceSampleMC as mcmc from pylab import * res, xmax, pmax = mcmc(LogLike, plist, m=1000, n =10000, ss=0.1) #oldres = np.load('whitenoise.npy') #res = np.vstack((oldres, res)) np.save('whitenoise', res) plist = xmax #print loglikelihood(plist) #plist = fmin(loglikelihood, plist) #plist = fmin_powell(loglikelihood, plist) sys.exit(0) p0 = plist[:np0] p1 = plist[np0:np1] p2 = plist[np1:np2] #p3 = plist[np2:np3] for i,p in enumerate(T2EFAC): md.__dict__[p] = np.abs(p0[i])
#print 'phi', np.log(phi) #print 'f', np.log10(f) return LogLike #print np.max(plist**4) #plist[16] = -16 #print np.max(plist**4) #print loglikelihood(plist) #sys.exit(0) LogLike = lambda x:loglikelihood(x) * -1. from testmcmc import SliceSampleMC as mcmc #from pylab import * oldres = np.load('rednoise.npy') res, xmax, pmax = mcmc(LogLike, plist, m=100, n = 10000, ss=0.1,progressbar=True) #sys.exit(0) res = np.array(res) res = np.vstack((oldres, res)) np.save('rednoise', res) plist = xmax #sys.exit(0) #plist = fmin(loglikelihood, plist) #plist = fmin_powell(loglikelihood, plist) p0 = plist[:np0] p1 = plist[np0:np1] p2 = plist[np1:np2] p3 = plist[np2:np3] for i,p in enumerate(T2EFAC): md.__dict__[p] = np.abs(p0[i]) for i,p in enumerate(T2EQUAD):
#print 'phi', np.log(phi) #print 'f', np.log10(f) return LogLike #print np.max(plist**4) #plist[16] = -16 #print np.max(plist**4) #print loglikelihood(plist) #sys.exit(0) LogLike = lambda x: loglikelihood(x) * -1. from testmcmc import SliceSampleMC as mcmc #from pylab import * res, xmax, pmax = mcmc(LogLike, plist, m=100, n=1000, ss=0.1, progressbar=True) #oldres = np.load('rednoise.npy') #res = np.vstack((oldres, res)) res = np.array(res) np.save('rednoise', res) plist = xmax #sys.exit(0) #plist = fmin(loglikelihood, plist) #plist = fmin_powell(loglikelihood, plist) p0 = plist[:np0] p1 = plist[np0:np1] p2 = plist[np1:np2] p3 = plist[np2:np3] for i, p in enumerate(T2EFAC): md.__dict__[p] = np.abs(p0[i]) for i, p in enumerate(T2EQUAD):