Python* API Reference for Intel® Data Analytics Acceleration Library 2018 Update 2

em_gmm_dense_batch.py

1 # file: em_gmm_dense_batch.py
2 #===============================================================================
3 # Copyright 2014-2018 Intel Corporation
4 # All Rights Reserved.
5 #
6 # If this software was obtained under the Intel Simplified Software License,
7 # the following terms apply:
8 #
9 # The source code, information and material ("Material") contained herein is
10 # owned by Intel Corporation or its suppliers or licensors, and title to such
11 # Material remains with Intel Corporation or its suppliers or licensors. The
12 # Material contains proprietary information of Intel or its suppliers and
13 # licensors. The Material is protected by worldwide copyright laws and treaty
14 # provisions. No part of the Material may be used, copied, reproduced,
15 # modified, published, uploaded, posted, transmitted, distributed or disclosed
16 # in any way without Intel's prior express written permission. No license under
17 # any patent, copyright or other intellectual property rights in the Material
18 # is granted to or conferred upon you, either expressly, by implication,
19 # inducement, estoppel or otherwise. Any license under such intellectual
20 # property rights must be express and approved by Intel in writing.
21 #
22 # Unless otherwise agreed by Intel in writing, you may not remove or alter this
23 # notice or any other notice embedded in Materials by Intel or Intel's
24 # suppliers or licensors in any way.
25 #
26 #
27 # If this software was obtained under the Apache License, Version 2.0 (the
28 # "License"), the following terms apply:
29 #
30 # You may not use this file except in compliance with the License. You may
31 # obtain a copy of the License at http://www.apache.org/licenses/LICENSE-2.0
32 #
33 #
34 # Unless required by applicable law or agreed to in writing, software
35 # distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
36 # WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
37 #
38 # See the License for the specific language governing permissions and
39 # limitations under the License.
40 #===============================================================================
41 
42 
44 
45 import os
46 import sys
47 
48 from daal.algorithms import em_gmm
49 from daal.data_management import FileDataSource, DataSourceIface
50 
51 utils_folder = os.path.realpath(os.path.abspath(os.path.dirname(os.path.dirname(__file__))))
52 if utils_folder not in sys.path:
53  sys.path.insert(0, utils_folder)
54 from utils import printNumericTable
55 
56 DAAL_PREFIX = os.path.join('..', 'data')
57 
58 # Input data set parameters
59 datasetFileName = os.path.join(DAAL_PREFIX, 'batch', 'em_gmm.csv')
60 nComponents = 2
61 
62 if __name__ == "__main__":
63 
64  # Initialize FileDataSource<CSVFeatureManager> to retrieve the input data from a .csv file
65  dataSource = FileDataSource(
66  datasetFileName,
67  DataSourceIface.doAllocateNumericTable,
68  DataSourceIface.doDictionaryFromContext
69  )
70  nFeatures = dataSource.getNumberOfColumns()
71 
72  # Retrieve the data from the input file
73  dataSource.loadDataBlock()
74 
75  # Create algorithm objects to initialize the EM algorithm for the GMM
76  # computing the number of components using the default method
77  initAlgorithm = em_gmm.init.Batch(nComponents)
78 
79  # Set an input data table for the initialization algorithm
80  initAlgorithm.input.set(em_gmm.init.data, dataSource.getNumericTable())
81 
82  # Compute initial values for the EM algorithm for the GMM with the default parameters
83  resultInit = initAlgorithm.compute()
84 
85  # Create algorithm objects for the EM algorithm for the GMM computing the number of components using the default method
86  algorithm = em_gmm.Batch(nComponents)
87 
88  # Set an input data table for the algorithm
89  algorithm.input.setTable(em_gmm.data, dataSource.getNumericTable())
90  algorithm.input.setValues(em_gmm.inputValues, resultInit)
91 
92  # Compute the results of the EM algorithm for the GMM with the default parameters
93  result = algorithm.compute()
94 
95  # Print the results
96  printNumericTable(result.getResult(em_gmm.weights), "Weights")
97  printNumericTable(result.getResult(em_gmm.means), "Means")
98  for i in range(nComponents):
99  printNumericTable(result.getCovariances(em_gmm.covariances, i), "Covariance")

For more complete information about compiler optimizations, see our Optimization Notice.