Python mv_normal_chol_like Examples

Example #1

File: step_methods.py Project: apatil/anopheles

 def hastings_factor(self):
     if self.cond_jumper.value is not None:
         for_factor = pm.mv_normal_chol_like(self.stochastic.value, self.cond_jumper.value.M_cond, self.cond_jumper.value.L_cond)
         back_factor = pm.mv_normal_chol_like(self.stochastic.last_value, self.cond_jumper.value.M_cond, self.cond_jumper.value.L_cond)
         return back_factor - for_factor
     else:
         return 0.

Example #2

File: step_methods.py Project: Bitburg-chef/pymc

    def propose(self):
        sig = pm.utils.value(self.stochastic.parents['sig'])
        mu = pm.utils.value(self.stochastic.parents['mu'])
        
        delta = pm.rmv_normal_chol(0*mu, sig)

        beta = np.minimum(1, self.proposal_sd * self.adaptive_scale_factor)
        bsig = beta*sig
        sb2 = np.sqrt(1-beta**2)
        self.stochastic.value = (self.stochastic.value - mu)*sb2+beta*delta+mu
        xp,x = self.stochastic.value, self.stochastic.last_value
        self._hastings_factor = pm.mv_normal_chol_like(x,(xp-mu)*sb2+mu,bsig) - pm.mv_normal_chol_like(xp,(x-mu)*sb2+mu,bsig)

Example #3

    def propose(self):
        sig = pm.utils.value(self.stochastic.parents['sig'])
        mu = pm.utils.value(self.stochastic.parents['mu'])
        
        delta = pm.rmv_normal_chol(0*mu, sig)

        beta = np.minimum(1, self.proposal_sd * self.adaptive_scale_factor)
        bsig = beta*sig
        sb2 = np.sqrt(1-beta**2)
        self.stochastic.value = (self.stochastic.value - mu)*sb2+beta*delta+mu
        xp,x = self.stochastic.value, self.stochastic.last_value
        self._hastings_factor = pm.mv_normal_chol_like(x,(xp-mu)*sb2+mu,bsig) - pm.mv_normal_chol_like(xp,(x-mu)*sb2+mu,bsig)

Example #4

File: pymc_test.py Project: xiangze/gpustats

def mixture_loglike(data, thetas, covs, labels):

    n = len(data)
    likes = pdfs.mvnpdf(data, thetas, covs)
    loglike = likes.ravel('F').take(labels * n + np.arange(n)).sum()

    if np.isnan(loglike):
        return -1e300

    return loglike

    if np.isnan(likes).any():
        loglike = 0.
        for j, (theta, cov) in enumerate(zip(thetas, covs)):
            this_data = data[labels == j]
            ch = chol(cov)
            loglike += pm.mv_normal_chol_like(this_data, theta, ch)

        return loglike

Example #5

File: pymc_test.py Project: drdangersimon/gpustats

def mixture_loglike(data, thetas, covs, labels):

    n = len(data)
    likes = pdfs.mvnpdf(data, thetas, covs)
    loglike = likes.ravel('F').take(labels * n + np.arange(n)).sum()

    if np.isnan(loglike):
        return -1e300

    return loglike

    if np.isnan(likes).any():
        loglike = 0.
        for j, (theta, cov) in enumerate(zip(thetas, covs)):
            this_data = data[labels == j]
            ch = chol(cov)
            loglike += pm.mv_normal_chol_like(this_data, theta, ch)

        return loglike

Example #6

File: pymc_test.py Project: xiangze/gpustats

def mixture_loglike2(data, thetas, taus, weights):

    n = len(data)

    covs = [inv(tau) for tau in taus]

    likes = pdfs.mvnpdf(data, thetas, covs)
    loglike = (likes * weights).sum()

    # loglike = likes.ravel('F').take(labels * n + np.arange(n)).sum()

    if np.isnan(loglike):
        st()
        return -1e300

    return loglike

    if np.isnan(likes).any():
        loglike = 0.
        for j, (theta, cov) in enumerate(zip(thetas, covs)):
            this_data = data[labels == j]
            loglike += pm.mv_normal_chol_like(this_data, theta, ch)

        return loglike

Example #7

File: pymc_test.py Project: drdangersimon/gpustats

def mixture_loglike2(data, thetas, taus, weights):

    n = len(data)

    covs = [inv(tau) for tau in taus]

    likes = pdfs.mvnpdf(data, thetas, covs)
    loglike = (likes * weights).sum()

    # loglike = likes.ravel('F').take(labels * n + np.arange(n)).sum()

    if np.isnan(loglike):
        st()
        return -1e300

    return loglike

    if np.isnan(likes).any():
        loglike = 0.
        for j, (theta, cov) in enumerate(zip(thetas, covs)):
            this_data = data[labels == j]
            loglike += pm.mv_normal_chol_like(this_data, theta, ch)

        return loglike