ó ‡ˆ\c@sĘdZddlZddlmZddlmZddlmZm Z m Z ddl m Z dd l mZmZdd lmZd d d dddgZied6ed6Zdefd„ƒYZdS(s5 Kernel Density Estimation ------------------------- i˙˙˙˙N(tgammainci(t BaseEstimator(t check_arraytcheck_random_statetcheck_consistent_length(t row_normsi(tBallTreetDTYPE(tKDTreetgaussianttophatt epanechnikovt exponentialtlineartcosinet ball_treetkd_treet KernelDensityc BsneZdZddddddeddd„ Zd„Zddd „Zd „Zdd „Z d dd „Z RS(sŤKernel Density Estimation Read more in the :ref:`User Guide `. Parameters ---------- bandwidth : float The bandwidth of the kernel. algorithm : string The tree algorithm to use. Valid options are ['kd_tree'|'ball_tree'|'auto']. Default is 'auto'. kernel : string The kernel to use. Valid kernels are ['gaussian'|'tophat'|'epanechnikov'|'exponential'|'linear'|'cosine'] Default is 'gaussian'. metric : string The distance metric to use. Note that not all metrics are valid with all algorithms. Refer to the documentation of :class:`BallTree` and :class:`KDTree` for a description of available algorithms. Note that the normalization of the density output is correct only for the Euclidean distance metric. Default is 'euclidean'. atol : float The desired absolute tolerance of the result. A larger tolerance will generally lead to faster execution. Default is 0. rtol : float The desired relative tolerance of the result. A larger tolerance will generally lead to faster execution. Default is 1E-8. breadth_first : boolean If true (default), use a breadth-first approach to the problem. Otherwise use a depth-first approach. leaf_size : int Specify the leaf size of the underlying tree. See :class:`BallTree` or :class:`KDTree` for details. Default is 40. metric_params : dict Additional parameters to be passed to the tree for use with the metric. For more information, see the documentation of :class:`BallTree` or :class:`KDTree`. gđ?tautoR t euclideanii(c CsŞ||_||_||_||_||_||_||_||_| |_|j |j|jƒ|dkr‚t dƒ‚n|t krŚt dj |ƒƒ‚ndS(Nisbandwidth must be positivesinvalid kernel: '{0}'( t algorithmt bandwidthtkerneltmetrictatoltrtolt breadth_firstt leaf_sizet metric_paramst_choose_algorithmt ValueErrort VALID_KERNELStformat( tselfRRRRRRRRR((s4lib/python2.7/site-packages/sklearn/neighbors/kde.pyt__init__Is           cCsĽ|dkrJ|tjkrdS|tjkr2dStdj|ƒƒ‚nW|tkrŒ|t|jkrˆtdjt||ƒƒ‚n|Stdj|ƒƒ‚dS(NRRRsinvalid metric: '{0}'sinvalid metric for {0}: '{1}'sinvalid algorithm: '{0}'(Rt valid_metricsRRR t TREE_DICT(R!RR((s4lib/python2.7/site-packages/sklearn/neighbors/kde.pyR`s   c Cs|j|j|jƒ}t|dddtƒ}|d k rżt|dddtdtƒ}|jdkrŽtdj |j d|j ƒƒ‚nt ||ƒ|j ƒdkrżtdƒ‚qżn|j }|d krÝi}nt||d |jd |jd |||_|S( s}Fit the Kernel Density model on the data. Parameters ---------- X : array_like, shape (n_samples, n_features) List of n_features-dimensional data points. Each row corresponds to a single data point. sample_weight : array_like, shape (n_samples,), optional List of sample weights attached to the data X. tordertCtdtypet ensure_2dis6the shape of sample_weight must be ({0},), but was {1}is'sample_weight must have positive valuesRRt sample_weightN(RRRRRtNonetFalsetndimRR tshapeRtminRR$Rttree_(R!tXtyR)Rtkwargs((s4lib/python2.7/site-packages/sklearn/neighbors/kde.pytfitts(           cCs˛t|dddtƒ}|jjd kr@|jjjd}n |jj}|j|}|jj |d|j d|j d|d|j d |j d tƒ}|tj|ƒ8}|S( sEvaluate the density model on the data. Parameters ---------- X : array_like, shape (n_samples, n_features) An array of points to query. Last dimension should match dimension of training data (n_features). Returns ------- density : ndarray, shape (n_samples,) The array of log(density) evaluations. These are normalized to be probability densities, so values will be low for high-dimensional data. R%R&R'ithRRRRt return_logN(RRR/R)R*tdataR-t sum_weightRtkernel_densityRRRRtTruetnptlog(R!R0tNtatol_Nt log_density((s4lib/python2.7/site-packages/sklearn/neighbors/kde.pyt score_samples–s   cCstj|j|ƒƒS(súCompute the total log probability density under the model. Parameters ---------- X : array_like, shape (n_samples, n_features) List of n_features-dimensional data points. Each row corresponds to a single data point. Returns ------- logprob : float Total log-likelihood of the data in X. This is normalized to be a probability density, so the value will be low for high-dimensional data. (R:tsumR?(R!R0R1((s4lib/python2.7/site-packages/sklearn/neighbors/kde.pytscoreľsic Cs“|jd krtƒ‚ntj|jjƒ}t|ƒ}|jddd|ƒ}|jjd kr†||j dj tj ƒ}n>tj tj|jjƒƒ}|d}tj|||ƒ}|jdkrótj|j|||jƒƒS|jdkr|j d} |jd|| fƒ} t| dtƒ} td| d| ƒd | |jtj| ƒ} ||| | d d …tjfSd S( sŇGenerate random samples from the model. Currently, this is implemented only for gaussian and tophat kernels. Parameters ---------- n_samples : int, optional Number of samples to generate. Defaults to 1. random_state : int, RandomState instance or None. default to None If int, random_state is the seed used by the random number generator; If RandomState instance, random_state is the random number generator; If None, the random number generator is the RandomState instance used by `np.random`. Returns ------- X : array_like, shape (n_samples, n_features) List of samples. R R iitsizei˙˙˙˙tsquaredgŕ?gđ?N(R R (RtNotImplementedErrorR:tasarrayR/R6RtuniformR)R*R-tastypetint64tcumsumt searchsortedt atleast_2dtnormalRRR9Rtsqrttnewaxis( R!t n_samplest random_stateR6trngtutit cumsum_weightR7tdimR0ts_sqt correction((s4lib/python2.7/site-packages/sklearn/neighbors/kde.pytsampleÇs$      3N( t__name__t __module__t__doc__R9R*R"RR3R?RARX(((s4lib/python2.7/site-packages/sklearn/neighbors/kde.pyRs/  "  (R[tnumpyR:t scipy.specialRtbaseRtutilsRRRt utils.extmathRRRRRRRR$R(((s4lib/python2.7/site-packages/sklearn/neighbors/kde.pyts