def getStringSummaryOfPrior(Prior):
''' Create string summarizing prior information
Returns
-------
s : str
'''
msg = 'Gaussian-Wishart joint prior on regression weights/prec\n'
msg += "Wishart on precision scalar\n"
msg += " pnu = %.3g\n" % (Prior.pnu)
msg += " ptau = %.3g\n" % (Prior.ptau)
msg += " mean = %.3g\n" % (E_d(pnu=Prior.pnu, ptau=Prior.ptau))
w_E_str = np2flatstr(Prior.w_E[:2], fmt='% .2g')
P_EE_str = np2flatstr(np.diag(Prior.P_EE)[:2], fmt='% .2g')
msg += "Gaussian on regression weight vector\n"
msg += " mean = %s\n" % w_E_str
msg += " diag[prec] = %s\n" % P_EE_str
return msg
def getStringSummaryOfPrior(Prior):
''' Get human-readable summary of specific prior
Returns
-------
s : str
'''
if Prior.D > 2:
sfx = ' ...'
else:
sfx = ''
msg = 'independent Gauss-Wishart prior on each dimension\n'
msg += "Wishart params \n"
msg += " nu = %.3g %s\n" % (Prior.nu, sfx)
msg += "beta = %s %s\n" % (np2flatstr(Prior.beta[:2]), sfx)
msg += "Expectations\n"
msg += 'E[ mean[k]] = \n%s%s\n' % (np2flatstr(Prior.m[:2]), sfx)
msg += 'E[ covar[k]] = \n'
S = E_CovMat(nu=Prior.nu, beta=Prior.beta[:2])
msg += str(S) + sfx
msg = msg.replace('\n', '\n ')
return msg