pyhrf.jde.nrl.gammagaussian module

class pyhrf.jde.nrl.gammagaussian.GamGaussMixtureParamsSampler(parameters=None, xmlHandler=None, xmlLabel=None, xmlComment=None)

Bases: pyhrf.jde.samplerbase.GibbsSamplerVariable

#TODO : comment

I_MEAN_CA = 0

I_VAR_CA = 1

I_VAR_CI = 2

NB_PARAMS = 3

PARAMS_NAMES = ['Shape_Activ', 'Scale_Activ', 'Var_Inactiv']

P_SAMPLE_FLAG = 'sampleFlag'

P_SCALE_CA_PR_ALPHA = 'scaleCAPrAlpha'

P_SCALE_CA_PR_BETA = 'scaleCAPrBeta'

P_SHAPE_CA_PR_MEAN = 'shapeCAPrMean'

P_VAL_INI = 'initialValue'

P_VAR_CI_PR_ALPHA = 'varCIPrAlpha'

P_VAR_CI_PR_BETA = 'varCIPrBeta'

checkAndSetInitValue(variables)

defaultParameters = {'initialValue': None, 'varCIPrBeta': 0.5, 'sampleFlag': 1, 'scaleCAPrAlpha': 2.5, 'varCIPrAlpha': 2.5, 'scaleCAPrBeta': 1.5, 'shapeCAPrMean': 10.0}

linkToData(dataInput)

sampleNextInternal(variables)

class pyhrf.jde.nrl.gammagaussian.InhomogeneousNRLSampler(parameters=None, xmlHandler=None, xmlLabel=None, xmlComment=None)

Bases: pyhrf.xmlio.Initable, pyhrf.jde.samplerbase.GibbsSamplerVariable

Class handling the Gibbs sampling of Neural Response Levels according to Salima Makni’s algorithm (IEEE SP 2005). Inherits the abstract class C{GibbsSamplerVariable}. #TODO : comment attributes

L_CA = 1

L_CI = 0

P_BETA = 'beta'

P_LABELS_COLORS = 'labelsColors'

P_LABELS_INI = 'labelsIni'

P_SAMPLE_FLAG = 'sampleFlag'

P_SAMPLE_LABELS = 'sampleLabels'

P_TRUE_LABELS = 'trueLabels'

P_VAL_INI = 'initialValue'

calcEnergy(voxIdx, label, cond)

checkAndSetInitValue(variables)

computeMean()

computeMeanClassApost(j, nrls, varXhj, rb)

computeVarYTilde(varXh)

computeVariablesApost(varCI, shapeCA, scaleCA, rb, varXh, varLambda)

countLabels()

defaultParameters = {'initialValue': None, 'sampleLabels': 1, 'labelsColors': array([ 0., 0.]), 'labelsIni': None, 'sampleFlag': 1, 'beta': 0.4}

finalizeSampling()

linkToData(dataInput)

sampleLabels(cond, varCI, varCA, meanCA)

sampleNextAlt(variables)

sampleNextInternal(variables)

samplingWarmUp(variables)

#TODO : comment