Python* API Reference for Intel® Data Analytics Acceleration Library 2020 Update 1

pca_cor_csr_distr.py

1 # file: pca_cor_csr_distr.py
2 #===============================================================================
3 # Copyright 2014-2020 Intel Corporation
4 #
5 # Licensed under the Apache License, Version 2.0 (the "License");
6 # you may not use this file except in compliance with the License.
7 # You may obtain a copy of the License at
8 #
9 # http://www.apache.org/licenses/LICENSE-2.0
10 #
11 # Unless required by applicable law or agreed to in writing, software
12 # distributed under the License is distributed on an "AS IS" BASIS,
13 # WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
14 # See the License for the specific language governing permissions and
15 # limitations under the License.
16 #===============================================================================
17 
18 
19 
20 
21 import os
22 import sys
23 
24 import numpy as np
25 
26 from daal import step1Local, step2Master
27 from daal.algorithms import covariance
28 from daal.algorithms import pca
29 
30 utils_folder = os.path.realpath(os.path.abspath(os.path.dirname(os.path.dirname(__file__))))
31 if utils_folder not in sys.path:
32  sys.path.insert(0, utils_folder)
33 from utils import printNumericTable, createSparseTable
34 
35 DAAL_PREFIX = os.path.join('..', 'data')
36 
37 # Input data set parameters
38 nBlocks = 4
39 datasetFileNames = [
40  os.path.join(DAAL_PREFIX, 'distributed', 'covcormoments_csr_1.csv'),
41  os.path.join(DAAL_PREFIX, 'distributed', 'covcormoments_csr_2.csv'),
42  os.path.join(DAAL_PREFIX, 'distributed', 'covcormoments_csr_3.csv'),
43  os.path.join(DAAL_PREFIX, 'distributed', 'covcormoments_csr_4.csv')
44 ]
45 
46 if __name__ == "__main__":
47 
48  # Create an algorithm for principal component analysis using the correlation method on the master node
49  masterAlgorithm = pca.Distributed(step2Master,fptype=np.float64)
50 
51  for i in range(nBlocks):
52  dataTable = createSparseTable(datasetFileNames[i])
53 
54  # Create algorithm objects to compute a variance-covariance matrix in the distributed processing mode using the default method
55  localAlgorithm = pca.Distributed(step1Local,fptype=np.float64)
56 
57  # Create an algorithm for principal component analysis using the correlation method on the local node
58  localAlgorithm.parameter.covariance = covariance.Distributed(step1Local, fptype=np.float64, method=covariance.fastCSR)
59 
60  # Set input objects for the algorithm
61  localAlgorithm.input.setDataset(pca.data, dataTable)
62 
63  # Compute partial estimates on local nodes
64  # Set local partial results as input for the master-node algorithm
65  masterAlgorithm.input.add(pca.partialResults, localAlgorithm.compute())
66 
67  # Use covariance algorithm for sparse data inside the PCA algorithm
68  masterAlgorithm.parameter.covariance = covariance.Distributed(step2Master, fptype=np.float64, method=covariance.fastCSR)
69 
70  # Merge and finalize PCA decomposition on the master node
71  masterAlgorithm.compute()
72 
73  result = masterAlgorithm.finalizeCompute()
74 
75  # Print the results
76  printNumericTable(result.get(pca.eigenvalues), "Eigenvalues:")
77  printNumericTable(result.get(pca.eigenvectors), "Eigenvectors:")

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