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

pca_metrics_dense_batch.py

1 # file: pca_metrics_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 import daal.algorithms.pca as pca
49 import daal.algorithms.pca.quality_metric_set as quality_metric_set
50 from daal.algorithms.pca.quality_metric import explained_variance
51 from daal.data_management import (
52  DataSourceIface, FileDataSource, HomogenNumericTable, MergedNumericTable,
53  NumericTableIface, BlockDescriptor, readWrite
54 )
55 
56 utils_folder = os.path.realpath(os.path.abspath(os.path.dirname(os.path.dirname(__file__))))
57 if utils_folder not in sys.path:
58  sys.path.insert(0, utils_folder)
59 from utils import printNumericTable
60 
61 datasetFileName = os.path.join('..', 'data', 'batch', 'pca_normalized.csv')
62 nVectors = 1000
63 nComponents = 5
64 
65 qmsResult = None
66 eigenData = None
67 
68 def trainModel():
69  global eigenData
70 
71  # Initialize FileDataSource to retrieve the input data from a .csv file
72  dataSource = FileDataSource(
73  datasetFileName,
74  DataSourceIface.doAllocateNumericTable,
75  DataSourceIface.doDictionaryFromContext
76  )
77 
78  # Retrieve the data from the input file
79  dataSource.loadDataBlock(nVectors)
80 
81  # Create an algorithm for principal component analysis using the SVD method
82  algorithm = pca.Batch(method=pca.svdDense)
83 
84  # Set the algorithm input data
85  algorithm.input.setDataset(pca.data, dataSource.getNumericTable())
86 
87  # Compute results of the PCA algorithm
88  result = algorithm.compute()
89  eigenData = result.get(pca.eigenvalues)
90 
91 def testPcaQuality():
92  global qmsResult
93 
94  # Create a quality metric set object to compute quality metrics of the PCA algorithm
95  qualityMetricSet = quality_metric_set.Batch(nComponents)
96  explainedVariances = explained_variance.Input.downCast(qualityMetricSet.getInputDataCollection().getInput(quality_metric_set.explainedVariancesMetrics))
97  explainedVariances.setInput(explained_variance.eigenvalues, eigenData)
98 
99  # Compute quality metrics
100  qualityMetricSet.compute()
101 
102  # Retrieve the quality metrics
103  qmsResult = qualityMetricSet.getResultCollection()
104 
105 def printResults():
106  print ("Quality metrics for PCA")
107  result = explained_variance.Result.downCast(qmsResult.getResult(quality_metric_set.explainedVariancesMetrics))
108  printNumericTable(result.getResult(explained_variance.explainedVariances), "Explained variances:")
109  printNumericTable(result.getResult(explained_variance.explainedVariancesRatios), "Explained variances ratios:")
110  printNumericTable(result.getResult(explained_variance.noiseVariance), "Noise variance:")
111 
112 if __name__ == "__main__":
113  trainModel()
114  testPcaQuality()
115  printResults()

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