diff options
Diffstat (limited to 'python/pyspark/mllib/clustering.py')
-rw-r--r-- | python/pyspark/mllib/clustering.py | 92 |
1 files changed, 88 insertions, 4 deletions
diff --git a/python/pyspark/mllib/clustering.py b/python/pyspark/mllib/clustering.py index 6b713aa393..f6b97abb17 100644 --- a/python/pyspark/mllib/clustering.py +++ b/python/pyspark/mllib/clustering.py @@ -15,19 +15,22 @@ # limitations under the License. # +from numpy import array + +from pyspark import RDD from pyspark import SparkContext from pyspark.mllib.common import callMLlibFunc, callJavaFunc -from pyspark.mllib.linalg import SparseVector, _convert_to_vector +from pyspark.mllib.linalg import DenseVector, SparseVector, _convert_to_vector +from pyspark.mllib.stat.distribution import MultivariateGaussian -__all__ = ['KMeansModel', 'KMeans'] +__all__ = ['KMeansModel', 'KMeans', 'GaussianMixtureModel', 'GaussianMixture'] class KMeansModel(object): """A clustering model derived from the k-means method. - >>> from numpy import array - >>> data = array([0.0,0.0, 1.0,1.0, 9.0,8.0, 8.0,9.0]).reshape(4,2) + >>> data = array([0.0,0.0, 1.0,1.0, 9.0,8.0, 8.0,9.0]).reshape(4, 2) >>> model = KMeans.train( ... sc.parallelize(data), 2, maxIterations=10, runs=30, initializationMode="random") >>> model.predict(array([0.0, 0.0])) == model.predict(array([1.0, 1.0])) @@ -86,6 +89,87 @@ class KMeans(object): return KMeansModel([c.toArray() for c in centers]) +class GaussianMixtureModel(object): + + """A clustering model derived from the Gaussian Mixture Model method. + + >>> clusterdata_1 = sc.parallelize(array([-0.1,-0.05,-0.01,-0.1, + ... 0.9,0.8,0.75,0.935, + ... -0.83,-0.68,-0.91,-0.76 ]).reshape(6, 2)) + >>> model = GaussianMixture.train(clusterdata_1, 3, convergenceTol=0.0001, + ... maxIterations=50, seed=10) + >>> labels = model.predict(clusterdata_1).collect() + >>> labels[0]==labels[1] + False + >>> labels[1]==labels[2] + True + >>> labels[4]==labels[5] + True + >>> clusterdata_2 = sc.parallelize(array([-5.1971, -2.5359, -3.8220, + ... -5.2211, -5.0602, 4.7118, + ... 6.8989, 3.4592, 4.6322, + ... 5.7048, 4.6567, 5.5026, + ... 4.5605, 5.2043, 6.2734]).reshape(5, 3)) + >>> model = GaussianMixture.train(clusterdata_2, 2, convergenceTol=0.0001, + ... maxIterations=150, seed=10) + >>> labels = model.predict(clusterdata_2).collect() + >>> labels[0]==labels[1]==labels[2] + True + >>> labels[3]==labels[4] + True + """ + + def __init__(self, weights, gaussians): + self.weights = weights + self.gaussians = gaussians + self.k = len(self.weights) + + def predict(self, x): + """ + Find the cluster to which the points in 'x' has maximum membership + in this model. + + :param x: RDD of data points. + :return: cluster_labels. RDD of cluster labels. + """ + if isinstance(x, RDD): + cluster_labels = self.predictSoft(x).map(lambda z: z.index(max(z))) + return cluster_labels + + def predictSoft(self, x): + """ + Find the membership of each point in 'x' to all mixture components. + + :param x: RDD of data points. + :return: membership_matrix. RDD of array of double values. + """ + if isinstance(x, RDD): + means, sigmas = zip(*[(g.mu, g.sigma) for g in self.gaussians]) + membership_matrix = callMLlibFunc("predictSoftGMM", x.map(_convert_to_vector), + self.weights, means, sigmas) + return membership_matrix + + +class GaussianMixture(object): + """ + Estimate model parameters with the expectation-maximization algorithm. + + :param data: RDD of data points + :param k: Number of components + :param convergenceTol: Threshold value to check the convergence criteria. Defaults to 1e-3 + :param maxIterations: Number of iterations. Default to 100 + :param seed: Random Seed + """ + @classmethod + def train(cls, rdd, k, convergenceTol=1e-3, maxIterations=100, seed=None): + """Train a Gaussian Mixture clustering model.""" + weight, mu, sigma = callMLlibFunc("trainGaussianMixture", + rdd.map(_convert_to_vector), k, + convergenceTol, maxIterations, seed) + mvg_obj = [MultivariateGaussian(mu[i], sigma[i]) for i in range(k)] + return GaussianMixtureModel(weight, mvg_obj) + + def _test(): import doctest globs = globals().copy() |