Method sample

numpy_ml/nonparametric/gp.py:223–258

Sample functions from the GP prior or posterior predictive distribution. Parameters ---------- X : :py:class:`ndarray ` of shape `(N, M)` The collection of datapoints to generate predictions on. Only used if `dist` = 'p

(self, X, n_samples=1, dist="posterior_predictive")

        return marginal_ll

    def sample(self, X, n_samples=1, dist="posterior_predictive"):
        """
        Sample functions from the GP prior or posterior predictive
        distribution.

        Parameters
        ----------
        X : :py:class:`ndarray <numpy.ndarray>` of shape `(N, M)`
            The collection of datapoints to generate predictions on. Only used if
            `dist` = 'posterior_predictive'.
        n_samples: int
            The number of samples to generate. Default is 1.
        dist : {"posterior_predictive", "prior"}
            The distribution to draw samples from. Default is
            "posterior_predictive".

        Returns
        -------
        samples : :py:class:`ndarray <numpy.ndarray>` of shape `(n_samples, O, N)`
            The generated samples for the points in `X`.
        """
        mvnorm = np.random.multivariate_normal

        if dist == "prior":
            mu = np.zeros((X.shape[0], 1))
            cov = self.kernel(X, X)
        elif dist == "posterior_predictive":
            mu, _, cov = self.predict(X, return_cov=True)
        else:
            raise ValueError("Unrecognized dist: '{}'".format(dist))

        if mu.ndim == 1:
            mu = mu[:, np.newaxis]

        samples = np.array([mvnorm(_mu, cov, size=n_samples) for _mu in mu.T])
        return samples.swapaxes(0, 1)