DPGMM.py 文件源码

def _step_E(self, points):
        """
        In this step the algorithm evaluates the responsibilities of each points in each cluster

        Parameters
        ----------
        points : an array (n_points,dim)

        Returns
        -------
        log_resp: an array (n_points,n_components)
            an array containing the logarithm of the responsibilities.
        log_prob_norm : an array (n_points,)
            logarithm of the probability of each sample in points

        """

        n_points,dim = points.shape

        log_gaussian = _log_normal_matrix(points,self.means,self.cov,'full',self.n_jobs)
        log_gaussian -= 0.5 * dim * np.log(self.nu)
        digamma_sum = np.sum(psi(.5 * (self.nu - np.arange(0, dim)[:,np.newaxis])),0)
        log_lambda = digamma_sum + dim * np.log(2)

        log_prob = self.log_weights + log_gaussian + 0.5 * (log_lambda - dim / self.beta)

        log_prob_norm = logsumexp(log_prob, axis=1)
        log_resp = log_prob - log_prob_norm[:,np.newaxis]

        return log_prob_norm,log_resp