2D Example#
Todo
Write this section.
import fesslix as flx
flx.load_engine()
import fesslix.gpr
import matplotlib.pyplot as plt
import numpy as np
Random Number Generator: MT19937 - initialized with rand()=391537905;
Random Number Generator: MT19937 - initialized with 1000 initial calls.
## load data
N_data_max = 300
dm_in = np.genfromtxt("../data/sample_01_dps_full.csv",delimiter=',')[:N_data_max,:]
dm_out = np.genfromtxt("../data/sample_01_out_full.csv",delimiter=',')[:N_data_max]
print(dm_in.shape, dm_out.shape)
#print(dm_in)
#print(dm_out)
(300, 2) (300,)
config_gp = { 'name':'gp_1', 'Ndim':2, 'mean_type':'universal', 'mean_polyo':0, 'kernel_lst':['gauss','gauss'] }
gp_1 = fesslix.gpr.gp(config_gp)
gp_1.get_name()
gp_1.condition_on(dm_in,dm_out,opt_noise=True)
print( gp_1.info() )
gp_1.optimize()
print( gp_1.info() )
gp_1.condition_on(dm_in,dm_out,init_pvec=False)
print( gp_1.info() )
{'type': 'singlegp', 'name': 'gp_1', 'mean': {'type': 'universal', 'para_vec': array([9.05771163]), 'normalizef': 1.0}, 'kernel': {'type': ['gauss', 'gauss'], 'para_vec': array([1.43225701, 1. , 1. ]), 'n_vec': array([ 1. , 301.60409606, 0.52678004]), 'kernel_sd': 1.4322570128931111}, 'noise': 0.3120960131089485, 'noise_log': ' =====================================\n 1D-optimization (minimization)\n =====================================\n»» LSE-results logl=-217.528 »» sd_obsv [24.61948 <- 0.717925] »» sd_Z [24.61825 <- 0.717889] »» sd_noise [0.246183 <- 7.178886e-3] \n *** best guess *** logl=-217.528 noise=7.178886e-3 sd_Z=0.717889\n»» LSE-results logl=-182.311 »» sd_obsv [0.588177 <- 1.015248] »» sd_Z [0.415904 <- 0.717889] »» sd_noise [0.415904 <- 0.717889] \n *** best guess *** logl=-182.311 noise=0.717889 sd_Z=0.717889\n»» LSE-results logl=-140.772 »» sd_obsv [2.894893 <- 0.721469] »» sd_Z [2.880526 <- 0.717889] »» sd_noise [0.288053 <- 0.0717889] \n *** best guess *** logl=-140.772 noise=0.0717889 sd_Z=0.717889\n fa=f(-4.93661)=217.528\n fb=f(-2.63403)=140.772\n fc=f(-0.331441)=182.311\n»» LSE-results logl=-150.252 »» sd_obsv [4.727501 <- 0.719098] »» sd_Z [4.719551 <- 0.717889] »» sd_noise [0.274052 <- 0.0416859] \n using brent method\n iter [lower, upper] xmin fmin err(est)\n 0\t-4.93661\t-0.331441\t-2.63403\t4.60517\n»» LSE-results logl=-158.871 »» sd_obsv [6.43511 <- 0.718507] »» sd_Z [6.429576 <- 0.717889] »» sd_noise [0.266819 <- 0.0297914] \n 1\t-3.51354\t-0.331441\t-2.63403\t3.18209\n»» LSE-results logl=-135.425 »» sd_obsv [1.347719 <- 0.738437] »» sd_Z [1.310216 <- 0.717889] »» sd_noise [0.315725 <- 0.172991] \n *** best guess *** logl=-135.425 noise=0.172991 sd_Z=0.717889\n 2\t-2.63403\t-0.331441\t-1.75452\t2.30259\n»» LSE-results logl=-135.323 »» sd_obsv [1.415463 <- 0.736181] »» sd_Z [1.380292 <- 0.717889] »» sd_noise [0.313576 <- 0.163091] \n *** best guess *** logl=-135.323 noise=0.163091 sd_Z=0.717889\n 3\t-2.63403\t-1.75452\t-1.81345\t0.879509\n»» LSE-results logl=-136.071 »» sd_obsv [1.85015 <- 0.727719] »» sd_Z [1.825157 <- 0.717889] »» sd_noise [0.303075 <- 0.119208] \n 4\t-2.12688\t-1.75452\t-1.81345\t0.372365\n»» LSE-results logl=-135.303 »» sd_obsv [1.474986 <- 0.734491] »» sd_Z [1.441645 <- 0.717889] »» sd_noise [0.311838 <- 0.155284] \n *** best guess *** logl=-135.303 noise=0.155284 sd_Z=0.717889\n 5\t-2.12688\t-1.81345\t-1.8625\t0.313433\n»» LSE-results logl=-135.302 »» sd_obsv [1.465021 <- 0.734757] »» sd_Z [1.431386 <- 0.717889] »» sd_noise [0.31212 <- 0.156539] \n *** best guess *** logl=-135.302 noise=0.156539 sd_Z=0.717889\n 6\t-1.8625\t-1.81345\t-1.85445\t0.0490487\n»» LSE-results logl=-135.302 »» sd_obsv [1.465943 <- 0.734733] »» sd_Z [1.432336 <- 0.717889] »» sd_noise [0.312094 <- 0.156422] \n *** best guess *** logl=-135.302 noise=0.156422 sd_Z=0.717889\n 7\t-1.8625\t-1.85445\t-1.8552\t0.00804605\n»» LSE-results logl=-135.302 »» sd_obsv [1.465868 <- 0.734735] »» sd_Z [1.432258 <- 0.717889] »» sd_noise [0.312096 <- 0.156431] \n *** best guess *** logl=-135.302 noise=0.156431 sd_Z=0.717889\n 8\t-1.8552\t-1.85445\t-1.85514\t0.00074725\n»» LSE-results logl=-135.302 »» sd_obsv [1.465866 <- 0.734735] »» sd_Z [1.432257 <- 0.717889] »» sd_noise [0.312096 <- 0.156432] \n *** best guess *** logl=-135.302 noise=0.156432 sd_Z=0.717889\n 9\t-1.85514\t-1.85445\t-1.85514\t0.000686356\n»» LSE-results logl=-135.302 »» sd_obsv [1.465866 <- 0.734735] »» sd_Z [1.432257 <- 0.717889] »» sd_noise [0.312096 <- 0.156432] \n 10\t-1.85514\t-1.85514\t-1.85514\t1.03279e-06\n Optimization converged\n»» LSE-results logl=-135.302 »» sd_obsv [1.465866 <- 0.734735] »» sd_Z [1.432257 <- 0.717889] »» sd_noise [0.312096 <- 0.156432] \n', 'opt_log': '', 'logl_obsv': -135.30211892323598, 'N_obsv': 300, 'obsv_up2date': True}
{'type': 'singlegp', 'name': 'gp_1', 'mean': {'type': 'universal', 'para_vec': array([9.07867198]), 'normalizef': 1.0}, 'kernel': {'type': ['gauss', 'gauss'], 'para_vec': array([1.43225701, 0.57230755, 1.1939226 ]), 'n_vec': array([ 1. , 301.60409606, 0.52678004]), 'kernel_sd': 1.4322570128931111}, 'noise': 0.3120960131089485, 'noise_log': '»» LSE-results logl=-131.875 »» sd_obsv [1.38838 <- 1.465866] »» sd_Z [1.356547 <- 1.432257] »» sd_noise [0.295598 <- 0.312096] \n', 'opt_log': ' flxGPProj::likeli_f_89 -135.302 ( 0, 0 ) no\n initial point estimate: -135.302 at ( 0, 0 ) dim=2\n flxGPProj::likeli_f_89 -135.302 ( 0, 0 ) no\n flxGPProj::likeli_f_89 -147.709 ( 1, 0 ) yes\n flxGPProj::likeli_f_89 -141.118 ( 0, 1 ) yes\n flxGPProj::likeli_f_89 -146.572 ( -1, 1 ) yes\n flxGPProj::likeli_f_89 -137.884 ( -0.5, 0.75 ) yes\n flxGPProj::likeli_f_89 -133.195 ( -0.5, -0.25 ) yes\n flxGPProj::likeli_f_89 -146.598 ( -0.75, -0.875 ) yes\n flxGPProj::likeli_f_89 -148.617 ( 0, -1 ) yes\n flxGPProj::likeli_f_89 -132.247 ( -0.375, 0.3125 ) yes\n flxGPProj::likeli_f_89 -133.779 ( -0.875, 0.0625 ) yes\n flxGPProj::likeli_f_89 -132.075 ( -0.65625, 0.046875 ) yes\n flxGPProj::likeli_f_89 -135.137 ( -0.53125, 0.609375 ) yes\n flxGPProj::likeli_f_89 -132.029 ( -0.507812, -0.0351562 ) yes\n flxGPProj::likeli_f_89 -135.908 ( -0.789062, -0.300781 ) yes\n flxGPProj::likeli_f_89 -131.996 ( -0.478516, 0.15918 ) yes\n flxGPProj::likeli_f_89 -132.541 ( -0.330078, 0.0771484 ) yes\n flxGPProj::likeli_f_89 -131.92 ( -0.574707, 0.0544434 ) yes\n flxGPProj::likeli_f_89 -131.881 ( -0.54541, 0.248779 ) yes\n flxGPProj::likeli_f_89 -132.351 ( -0.564209, 0.390747 ) yes\n flxGPProj::likeli_f_89 -131.986 ( -0.641602, 0.144043 ) yes\n flxGPProj::likeli_f_89 -131.905 ( -0.60083, 0.147827 ) yes\n flxGPProj::likeli_f_89 -132.083 ( -0.571533, 0.342163 ) yes\n flxGPProj::likeli_f_89 -131.897 ( -0.573914, 0.126373 ) yes\n flxGPProj::likeli_f_89 -131.9 ( -0.518494, 0.227325 ) yes\n flxGPProj::likeli_f_89 -131.878 ( -0.539078, 0.207451 ) yes\n flxGPProj::likeli_f_89 -132.074 ( -0.510574, 0.329857 ) yes\n flxGPProj::likeli_f_89 -131.875 ( -0.558079, 0.177244 ) yes\n MLE: -131.875 at ( -0.558079, 0.177244 ) :: niter = 27, dim = 2\n flxGPProj::likeli_f_89 -131.875 ( -0.558079, 0.177244 ) no\n', 'logl_obsv': -131.87529828420332, 'N_obsv': 300, 'obsv_up2date': True}
{'type': 'singlegp', 'name': 'gp_1', 'mean': {'type': 'universal', 'para_vec': array([9.08218271]), 'normalizef': 1.0}, 'kernel': {'type': ['gauss', 'gauss'], 'para_vec': array([1.3320121 , 0.57230755, 1.1939226 ]), 'n_vec': array([ 1. , 301.60409606, 0.52678004]), 'kernel_sd': 1.3320120953441206}, 'noise': 0.29646855025884833, 'noise_log': ' =====================================\n 1D-optimization (minimization)\n =====================================\n»» LSE-results logl=-203.305 »» sd_obsv [11.19383 <- 1.432597] »» sd_Z [11.19117 <- 1.432257] »» sd_noise [0.243861 <- 0.0312096] \n *** best guess *** logl=-203.305 noise=0.0312096 sd_Z=1.432257\n»» LSE-results logl=-253.821 »» sd_obsv [0.599269 <- 3.433912] »» sd_Z [0.24995 <- 1.432257] »» sd_noise [0.544654 <- 3.12096] \n»» LSE-results logl=-131.875 »» sd_obsv [1.38838 <- 1.465866] »» sd_Z [1.356547 <- 1.432257] »» sd_noise [0.295598 <- 0.312096] \n *** best guess *** logl=-131.875 noise=0.312096 sd_Z=1.432257\n fa=f(-3.46703)=203.305\n fb=f(-1.16444)=131.875\n fc=f(1.13814)=253.821\n»» LSE-results logl=-137.063 »» sd_obsv [2.20581 <- 1.443677] »» sd_Z [2.188362 <- 1.432257] »» sd_noise [0.276897 <- 0.181226] \n using brent method\n iter [lower, upper] xmin fmin err(est)\n 0\t-3.46703\t1.13814\t-1.16444\t4.60517\n»» LSE-results logl=-144.369 »» sd_obsv [2.975207 <- 1.438101] »» sd_Z [2.963116 <- 1.432257] »» sd_noise [0.267948 <- 0.129516] \n 1\t-2.04395\t1.13814\t-1.16444\t3.18209\n»» LSE-results logl=-131.915 »» sd_obsv [1.30946 <- 1.471018] »» sd_Z [1.274956 <- 1.432257] »» sd_noise [0.298617 <- 0.335459] \n 2\t-2.04395\t-1.09225\t-1.16444\t0.951699\n»» LSE-results logl=-131.867 »» sd_obsv [1.366893 <- 1.467162] »» sd_Z [1.334374 <- 1.432257] »» sd_noise [0.296383 <- 0.318125] \n *** best guess *** logl=-131.867 noise=0.318125 sd_Z=1.432257\n 3\t-1.16444\t-1.09225\t-1.14531\t0.0721897\n»» LSE-results logl=-131.867 »» sd_obsv [1.36476 <- 1.467294] »» sd_Z [1.332171 <- 1.432257] »» sd_noise [0.296463 <- 0.318736] \n *** best guess *** logl=-131.867 noise=0.318736 sd_Z=1.432257\n 4\t-1.14531\t-1.09225\t-1.14339\t0.0530574\n»» LSE-results logl=-131.867 »» sd_obsv [1.3646 <- 1.467304] »» sd_Z [1.332006 <- 1.432257] »» sd_noise [0.296469 <- 0.318782] \n *** best guess *** logl=-131.867 noise=0.318782 sd_Z=1.432257\n 5\t-1.14339\t-1.09225\t-1.14325\t0.0511374\n»» LSE-results logl=-131.867 »» sd_obsv [1.364606 <- 1.467304] »» sd_Z [1.332012 <- 1.432257] »» sd_noise [0.296469 <- 0.31878] \n *** best guess *** logl=-131.867 noise=0.31878 sd_Z=1.432257\n 6\t-1.14339\t-1.14325\t-1.14325\t0.000144486\n Optimization converged\n»» LSE-results logl=-131.867 »» sd_obsv [1.364606 <- 1.467304] »» sd_Z [1.332012 <- 1.432257] »» sd_noise [0.296469 <- 0.31878] \n', 'opt_log': ' flxGPProj::likeli_f_89 -135.302 ( 0, 0 ) no\n initial point estimate: -135.302 at ( 0, 0 ) dim=2\n flxGPProj::likeli_f_89 -135.302 ( 0, 0 ) no\n flxGPProj::likeli_f_89 -147.709 ( 1, 0 ) yes\n flxGPProj::likeli_f_89 -141.118 ( 0, 1 ) yes\n flxGPProj::likeli_f_89 -146.572 ( -1, 1 ) yes\n flxGPProj::likeli_f_89 -137.884 ( -0.5, 0.75 ) yes\n flxGPProj::likeli_f_89 -133.195 ( -0.5, -0.25 ) yes\n flxGPProj::likeli_f_89 -146.598 ( -0.75, -0.875 ) yes\n flxGPProj::likeli_f_89 -148.617 ( 0, -1 ) yes\n flxGPProj::likeli_f_89 -132.247 ( -0.375, 0.3125 ) yes\n flxGPProj::likeli_f_89 -133.779 ( -0.875, 0.0625 ) yes\n flxGPProj::likeli_f_89 -132.075 ( -0.65625, 0.046875 ) yes\n flxGPProj::likeli_f_89 -135.137 ( -0.53125, 0.609375 ) yes\n flxGPProj::likeli_f_89 -132.029 ( -0.507812, -0.0351562 ) yes\n flxGPProj::likeli_f_89 -135.908 ( -0.789062, -0.300781 ) yes\n flxGPProj::likeli_f_89 -131.996 ( -0.478516, 0.15918 ) yes\n flxGPProj::likeli_f_89 -132.541 ( -0.330078, 0.0771484 ) yes\n flxGPProj::likeli_f_89 -131.92 ( -0.574707, 0.0544434 ) yes\n flxGPProj::likeli_f_89 -131.881 ( -0.54541, 0.248779 ) yes\n flxGPProj::likeli_f_89 -132.351 ( -0.564209, 0.390747 ) yes\n flxGPProj::likeli_f_89 -131.986 ( -0.641602, 0.144043 ) yes\n flxGPProj::likeli_f_89 -131.905 ( -0.60083, 0.147827 ) yes\n flxGPProj::likeli_f_89 -132.083 ( -0.571533, 0.342163 ) yes\n flxGPProj::likeli_f_89 -131.897 ( -0.573914, 0.126373 ) yes\n flxGPProj::likeli_f_89 -131.9 ( -0.518494, 0.227325 ) yes\n flxGPProj::likeli_f_89 -131.878 ( -0.539078, 0.207451 ) yes\n flxGPProj::likeli_f_89 -132.074 ( -0.510574, 0.329857 ) yes\n flxGPProj::likeli_f_89 -131.875 ( -0.558079, 0.177244 ) yes\n MLE: -131.875 at ( -0.558079, 0.177244 ) :: niter = 27, dim = 2\n flxGPProj::likeli_f_89 -131.875 ( -0.558079, 0.177244 ) no\n', 'logl_obsv': -131.86709066589054, 'N_obsv': 300, 'obsv_up2date': True}
config_gp = { 'name':'gp_2', 'Ndim':2, 'mean_type':'universal', 'mean_polyo':0, 'kernel_lst':['gauss','gauss'], 'useLSE':False }
gp_2 = fesslix.gpr.gp(config_gp)
gp_2.noise_white(0.5)
print( gp_2.info() )
gp_2.condition_on(dm_in,dm_out,opt_noise=True) #,opt_noise=False
print( gp_2.info() )
gp_2.optimize()
print( gp_2.info() )
{'type': 'singlegp', 'name': 'gp_2', 'mean': {'type': 'universal', 'para_vec': array([1.]), 'normalizef': 1.0}, 'kernel': {'type': ['gauss', 'gauss'], 'para_vec': array([1., 1., 1.]), 'n_vec': array([1., 1., 1.]), 'kernel_sd': 1.0}, 'noise': 0.5, 'noise_log': '', 'opt_log': '', 'N_obsv': 0, 'obsv_up2date': False}
{'type': 'singlegp', 'name': 'gp_2', 'mean': {'type': 'universal', 'para_vec': array([1.]), 'normalizef': 10.257333333333333}, 'kernel': {'type': ['gauss', 'gauss'], 'para_vec': array([1., 1., 1.]), 'n_vec': array([ 0.71788864, 301.60409606, 0.52678004]), 'kernel_sd': 0.7178886421751652}, 'noise': 0.33117755511473074, 'noise_log': ' =====================================\n 1D-optimization (minimization)\n =====================================\n *** best guess *** logl=-1.75414e5 noise=7.178886e-3 sd_Z=0.717889\n *** best guess *** logl=-252.34 noise=0.717889 sd_Z=0.717889\n fa=f(-4.93661)=175414\n fb=f(-2.63403)=2010.71\n fc=f(-0.331441)=252.34\n bracketing the mimimum\n 0: fc=f(3.39422)=1294.12\n *** best guess *** logl=-252.34 noise=0.717889 sd_Z=0.717889\n using brent method\n iter [lower, upper] xmin fmin err(est)\n 0\t-2.63403\t3.39422\t-0.331441\t6.02825\n 1\t-2.63403\t1.09163\t-0.331441\t3.72566\n *** best guess *** logl=-157.522 noise=0.297914 sd_Z=0.717889\n 2\t-2.63403\t-0.331441\t-1.21095\t2.30259\n 3\t-1.25479\t-0.331441\t-1.21095\t0.923354\n 4\t-1.25479\t-0.875007\t-1.21095\t0.379788\n *** best guess *** logl=-154.657 noise=0.338703 sd_Z=0.717889\n 5\t-1.21095\t-0.875007\t-1.08263\t0.335943\n *** best guess *** logl=-154.535 noise=0.330171 sd_Z=0.717889\n 6\t-1.21095\t-1.08263\t-1.10815\t0.128319\n *** best guess *** logl=-154.533 noise=0.331391 sd_Z=0.717889\n 7\t-1.10815\t-1.08263\t-1.10446\t0.025514\n *** best guess *** logl=-154.533 noise=0.331183 sd_Z=0.717889\n 8\t-1.10815\t-1.10446\t-1.10508\t0.0036894\n *** best guess *** logl=-154.533 noise=0.331178 sd_Z=0.717889\n 9\t-1.10815\t-1.10508\t-1.1051\t0.00306199\n *** best guess *** logl=-154.533 noise=0.331178 sd_Z=0.717889\n 10\t-1.10815\t-1.1051\t-1.1051\t0.00304531\n *** best guess *** logl=-154.533 noise=0.331178 sd_Z=0.717889\n 11\t-1.1051\t-1.1051\t-1.1051\t6.21451e-07\n Optimization converged\n', 'opt_log': '', 'logl_obsv': -154.5327961734892, 'N_obsv': 300, 'obsv_up2date': True}
{'type': 'singlegp', 'name': 'gp_2', 'mean': {'type': 'universal', 'para_vec': array([0.87866234]), 'normalizef': 10.257333333333333}, 'kernel': {'type': ['gauss', 'gauss'], 'para_vec': array([1.85394249, 0.80040195, 1.38794787]), 'n_vec': array([ 0.71788864, 301.60409606, 0.52678004]), 'kernel_sd': 1.3309242560578016}, 'noise': 0.33117755511473074, 'noise_log': '', 'opt_log': ' flxGPProj::likeli_f_89 -154.533 ( 1, 0, 0, 0 ) no\n initial point estimate: -154.533 at ( 1, 0, 0, 0 ) dim=4\n flxGPProj::likeli_f_89 -154.533 ( 1, 0, 0, 0 ) no\n flxGPProj::likeli_f_89 -1273.36 ( 2, 0, 0, 0 ) yes\n flxGPProj::likeli_f_89 -140.294 ( 1, 1, 0, 0 ) yes\n flxGPProj::likeli_f_89 -150.568 ( 1, 1, 1, 0 ) yes\n flxGPProj::likeli_f_89 -142.83 ( 1, 1, 0, 1 ) yes\n flxGPProj::likeli_f_89 -170.223 ( 0, 1.5, 0.5, 0.5 ) yes\n flxGPProj::likeli_f_89 -151.702 ( 0.5, 1.125, 0.375, 0.375 ) yes\n flxGPProj::likeli_f_89 -150.447 ( 0.75, 2.0625, 0.6875, 0.6875 ) yes\n flxGPProj::likeli_f_89 -155.681 ( 1.375, 1.40625, 0.46875, 0.46875 ) yes\n flxGPProj::likeli_f_89 -142.969 ( 0.71875, 1.195312, 0.398438, 0.398438 ) yes\n flxGPProj::likeli_f_89 -150.681 ( 0.734375, 1.628906, -0.457031, 1.042969 ) yes\n flxGPProj::likeli_f_89 -145.859 ( 0.933594, 1.157227, 0.635742, 0.260742 ) yes\n flxGPProj::likeli_f_89 -165.13 ( 1.076172, 0.11377, -0.17041, 0.14209 ) yes\n flxGPProj::likeli_f_89 -144.689 ( 0.831543, 1.575317, 0.473022, 0.551147 ) yes\n flxGPProj::likeli_f_89 -138.939 ( 0.841553, 1.228088, -0.200012, 0.71405 ) yes\n flxGPProj::likeli_f_89 -146.032 ( 0.795532, 1.263519, -0.617889, 0.940704 ) yes\n flxGPProj::likeli_f_89 -136.053 ( 0.948608, 0.636383, -0.37381, 0.505096 ) yes\n flxGPProj::likeli_f_89 -154.804 ( 1.007141, 0.166916, -0.797226, 0.482071 ) yes\n flxGPProj::likeli_f_89 -162.995 ( 1.176331, 0.736923, -0.685349, 0.711136 ) yes\n flxGPProj::likeli_f_89 -136.911 ( 0.833145, 1.080715, 0.127491, 0.476612 ) yes\n flxGPProj::likeli_f_89 -139.93 ( 0.811653, 0.972593, -0.223166, -0.152121 ) yes\n flxGPProj::likeli_f_89 -141.546 ( 0.71748, 0.95889, -0.334748, 0.771819 ) yes\n flxGPProj::likeli_f_89 -136.762 ( 0.92937, 0.989722, -0.0836871, 0.192955 ) yes\n flxGPProj::likeli_f_89 -143.712 ( 0.964685, 0.994861, -0.0418435, 1.096477 ) yes\n flxGPProj::likeli_f_89 -136.442 ( 0.849911, 0.97816, -0.177835, 0.160029 ) yes\n flxGPProj::likeli_f_89 -136.718 ( 0.938965, 0.614402, -0.0539082, -0.0467042 ) yes\n flxGPProj::likeli_f_89 -140.792 ( 1.000282, 0.528619, -0.472111, -0.0709241 ) yes\n flxGPProj::likeli_f_89 -135.101 ( 0.874929, 0.942691, -0.0224095, 0.339728 ) yes\n flxGPProj::likeli_f_89 -133.366 ( 0.876837, 0.596096, -0.230294, 0.28612 ) yes\n flxGPProj::likeli_f_89 -134.874 ( 0.85057, 0.399282, -0.303598, 0.332702 ) yes\n flxGPProj::likeli_f_89 -137.521 ( 0.836178, 0.962263, -0.348266, 0.692191 ) yes\n flxGPProj::likeli_f_89 -134.699 ( 0.913268, 0.701367, -0.127498, 0.13802 ) yes\n flxGPProj::likeli_f_89 -136.702 ( 0.95691, 0.460108, -0.199171, 0.474453 ) yes\n flxGPProj::likeli_f_89 -134.593 ( 0.876661, 0.848647, -0.183169, 0.238635 ) yes\n flxGPProj::likeli_f_89 -138.538 ( 0.822239, 0.908018, 0.0921246, -3.845265e-3 ) yes\n flxGPProj::likeli_f_89 -133.962 ( 0.917016, 0.704292, -0.257326, 0.377861 ) yes\n flxGPProj::likeli_f_89 -134.44 ( 0.916961, 0.48251, -0.376734, 0.18059 ) yes\n flxGPProj::likeli_f_89 -133.894 ( 0.88047, 0.614405, -0.396264, 0.403583 ) yes\n flxGPProj::likeli_f_89 -136.214 ( 0.918981, 0.350004, -0.44714, 0.385442 ) yes\n flxGPProj::likeli_f_89 -133.728 ( 0.887241, 0.723986, -0.249162, 0.275337 ) yes\n flxGPProj::likeli_f_89 -134.229 ( 0.86382, 0.83688, -0.189789, 0.490861 ) yes\n flxGPProj::likeli_f_89 -133.681 ( 0.877106, 0.748287, -0.236525, 0.413293 ) yes\n flxGPProj::likeli_f_89 -133.816 ( 0.84381, 0.637096, -0.298796, 0.311305 ) yes\n flxGPProj::likeli_f_89 -134.059 ( 0.862027, 0.738327, -0.111125, 0.239444 ) yes\n flxGPProj::likeli_f_89 -133.55 ( 0.875859, 0.645386, -0.324979, 0.362548 ) yes\n flxGPProj::likeli_f_89 -133.878 ( 0.914711, 0.719782, -0.221684, 0.357344 ) yes\n flxGPProj::likeli_f_89 -133.512 ( 0.861535, 0.657767, -0.279518, 0.322815 ) yes\n flxGPProj::likeli_f_89 -133.748 ( 0.858427, 0.599781, -0.286497, 0.417051 ) yes\n flxGPProj::likeli_f_89 -133.506 ( 0.880038, 0.692935, -0.258496, 0.310765 ) yes\n flxGPProj::likeli_f_89 -133.678 ( 0.870029, 0.547805, -0.310118, 0.227831 ) yes\n flxGPProj::likeli_f_89 -133.477 ( 0.871798, 0.597925, -0.29172, 0.274197 ) yes\n flxGPProj::likeli_f_89 -133.58 ( 0.869245, 0.626976, -0.205035, 0.2344 ) yes\n flxGPProj::likeli_f_89 -133.444 ( 0.874205, 0.640783, -0.294993, 0.330511 ) yes\n flxGPProj::likeli_f_89 -133.429 ( 0.889903, 0.606103, -0.258233, 0.277982 ) yes\n flxGPProj::likeli_f_89 -133.52 ( 0.876334, 0.527518, -0.279125, 0.273639 ) yes\n flxGPProj::likeli_f_89 -133.42 ( 0.879112, 0.651581, -0.263653, 0.301484 ) yes\n flxGPProj::likeli_f_89 -133.382 ( 0.888231, 0.649356, -0.231867, 0.323852 ) yes\n flxGPProj::likeli_f_89 -133.475 ( 0.892836, 0.610784, -0.19703, 0.264207 ) yes\n flxGPProj::likeli_f_89 -133.388 ( 0.878863, 0.633284, -0.270502, 0.313935 ) yes\n flxGPProj::likeli_f_89 -133.378 ( 0.871618, 0.659056, -0.239925, 0.334714 ) yes\n flxGPProj::likeli_f_89 -133.317 ( 0.878662, 0.617314, -0.222641, 0.327826 ) yes\n flxGPProj::likeli_f_89 -133.32 ( 0.878438, 0.600181, -0.202135, 0.340998 ) yes\n flxGPProj::likeli_f_89 -133.325 ( 0.878811, 0.627627, -0.191861, 0.32232 ) yes\n flxGPProj::likeli_f_89 -133.392 ( 0.864733, 0.600691, -0.210494, 0.311638 ) yes\n flxGPProj::likeli_f_89 -133.34 ( 0.882356, 0.63719, -0.226523, 0.320798 ) yes\n flxGPProj::likeli_f_89 -133.38 ( 0.886715, 0.580058, -0.195735, 0.293819 ) yes\n flxGPProj::likeli_f_89 -133.339 ( 0.875392, 0.639306, -0.228877, 0.32449 ) yes\n flxGPProj::likeli_f_89 -133.358 ( 0.880774, 0.664623, -0.204657, 0.361598 ) yes\n flxGPProj::likeli_f_89 -133.331 ( 0.87979, 0.647491, -0.211067, 0.342728 ) yes\n flxGPProj::likeli_f_89 -133.32 ( 0.873971, 0.62868, -0.2007, 0.337884 ) yes\n flxGPProj::likeli_f_89 -133.322 ( 0.880225, 0.62125, -0.184257, 0.34089 ) yes\n flxGPProj::likeli_f_89 -133.332 ( 0.876045, 0.599945, -0.188663, 0.321732 ) yes\n flxGPProj::likeli_f_89 -133.319 ( 0.878853, 0.635605, -0.205466, 0.337479 ) yes\n MLE: -133.317 at ( 0.878662, 0.617314, -0.222641, 0.327826 ) :: niter = 73, dim = 4\n flxGPProj::likeli_f_89 -133.317 ( 0.878662, 0.617314, -0.222641, 0.327826 ) yes\n', 'logl_obsv': -133.3174658738168, 'N_obsv': 300, 'obsv_up2date': True}
print( gp_1.predict([3500.,7.15],type="trend",predict_noise=False), gp_1.predict([3500.,7.15],type="trend",predict_noise=True) )
print( gp_1.predict([3500.,7.15],type="mean",predict_noise=False), gp_1.predict([3500.,7.15],type="mean",predict_noise=True) )
print( gp_1.predict([3500.,7.15],type="sd",predict_noise=False), gp_1.predict([3500.,7.15],type="sd",predict_noise=True) )
9.082182705424747 9.082182705424747
9.834747448394019 9.834747448394019
0.4794831779879559 0.5637355047058976