Python* API Reference for Intel® Data Analytics Acceleration Library 2018 Update 1

mn_naive_bayes_dense_distr.py

1 #===============================================================================
2 # Copyright 2014-2017 Intel Corporation
3 # All Rights Reserved.
4 #
5 # If this software was obtained under the Intel Simplified Software License,
6 # the following terms apply:
7 #
8 # The source code, information and material ("Material") contained herein is
9 # owned by Intel Corporation or its suppliers or licensors, and title to such
10 # Material remains with Intel Corporation or its suppliers or licensors. The
11 # Material contains proprietary information of Intel or its suppliers and
12 # licensors. The Material is protected by worldwide copyright laws and treaty
13 # provisions. No part of the Material may be used, copied, reproduced,
14 # modified, published, uploaded, posted, transmitted, distributed or disclosed
15 # in any way without Intel's prior express written permission. No license under
16 # any patent, copyright or other intellectual property rights in the Material
17 # is granted to or conferred upon you, either expressly, by implication,
18 # inducement, estoppel or otherwise. Any license under such intellectual
19 # property rights must be express and approved by Intel in writing.
20 #
21 # Unless otherwise agreed by Intel in writing, you may not remove or alter this
22 # notice or any other notice embedded in Materials by Intel or Intel's
23 # suppliers or licensors in any way.
24 #
25 #
26 # If this software was obtained under the Apache License, Version 2.0 (the
27 # "License"), the following terms apply:
28 #
29 # You may not use this file except in compliance with the License. You may
30 # obtain a copy of the License at http://www.apache.org/licenses/LICENSE-2.0
31 #
32 #
33 # Unless required by applicable law or agreed to in writing, software
34 # distributed under the License is distributed on an "AS IS" BASIS, WITHOUT
35 # WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
36 #
37 # See the License for the specific language governing permissions and
38 # limitations under the License.
39 #===============================================================================
40 
41 ## <a name="DAAL-EXAMPLE-PY-MULTINOMIAL_NAIVE_BAYES_DENSE_DISTRIBUTED"></a>
42 ## \example mn_naive_bayes_dense_distr.py
43 
44 import os
45 import sys
46 
47 from daal import step1Local, step2Master
48 from daal.algorithms.multinomial_naive_bayes import prediction, training
49 from daal.algorithms import classifier
50 from daal.data_management import (
51  FileDataSource, DataSourceIface, NumericTableIface, HomogenNumericTable, MergedNumericTable
52 )
53 
54 utils_folder = os.path.realpath(os.path.abspath(os.path.dirname(os.path.dirname(__file__))))
55 if utils_folder not in sys.path:
56  sys.path.insert(0, utils_folder)
57 from utils import printNumericTables
58 
59 DAAL_PREFIX = os.path.join('..', 'data')
60 
61 # Input data set parameters
62 trainDatasetFileNames = [
63  os.path.join(DAAL_PREFIX, 'batch', 'naivebayes_train_dense.csv'),
64  os.path.join(DAAL_PREFIX, 'batch', 'naivebayes_train_dense.csv'),
65  os.path.join(DAAL_PREFIX, 'batch', 'naivebayes_train_dense.csv'),
66  os.path.join(DAAL_PREFIX, 'batch', 'naivebayes_train_dense.csv')
67 ]
68 
69 testDatasetFileName = os.path.join(DAAL_PREFIX, 'batch', 'naivebayes_test_dense.csv')
70 
71 nFeatures = 20
72 nClasses = 20
73 nBlocks = 4
74 
75 trainingResult = None
76 predictionResult = None
77 testGroundTruth = None
78 
79 
80 def trainModel():
81  global trainingResult
82 
83  masterAlgorithm = training.Distributed(step2Master, nClasses)
84 
85  for i in range(nBlocks):
86  # Initialize FileDataSource<CSVFeatureManager> to retrieve the input data from a .csv file
87  trainDataSource = FileDataSource(
88  trainDatasetFileNames[i], DataSourceIface.notAllocateNumericTable,
89  DataSourceIface.doDictionaryFromContext
90  )
91  # Create Numeric Tables for training data and labels
92  trainData = HomogenNumericTable(nFeatures, 0, NumericTableIface.doNotAllocate)
93  trainGroundTruth = HomogenNumericTable(1, 0, NumericTableIface.doNotAllocate)
94  mergedData = MergedNumericTable(trainData, trainGroundTruth)
95 
96  # Retrieve the data from the input file
97  trainDataSource.loadDataBlock(mergedData)
98 
99  # Create an algorithm object to train the Naive Bayes model on the local-node data
100  localAlgorithm = training.Distributed(step1Local, nClasses)
101 
102  # Pass a training data set and dependent values to the algorithm
103  localAlgorithm.input.set(classifier.training.data, trainData)
104  localAlgorithm.input.set(classifier.training.labels, trainGroundTruth)
105 
106  # Build the Naive Bayes model on the local node and
107  # Set the local Naive Bayes model as input for the master-node algorithm
108  masterAlgorithm.input.add(training.partialModels, localAlgorithm.compute())
109 
110  # Merge and finalize the Naive Bayes model on the master node
111  masterAlgorithm.compute()
112  trainingResult = masterAlgorithm.finalizeCompute() # Retrieve the algorithm results
113 
114 
115 def testModel():
116  global predictionResult, testGroundTruth
117 
118  # Initialize FileDataSource<CSVFeatureManager> to retrieve the test data from a .csv file
119  testDataSource = FileDataSource(
120  testDatasetFileName, DataSourceIface.notAllocateNumericTable,
121  DataSourceIface.doDictionaryFromContext
122  )
123 
124  # Create Numeric Tables for testing data and labels
125  testData = HomogenNumericTable(nFeatures, 0, NumericTableIface.doNotAllocate)
126  testGroundTruth = HomogenNumericTable(1, 0, NumericTableIface.doNotAllocate)
127  mergedData = MergedNumericTable(testData, testGroundTruth)
128 
129  # Retrieve the data from input file
130  testDataSource.loadDataBlock(mergedData)
131 
132  # Create an algorithm object to predict Naive Bayes values
133  algorithm = prediction.Batch(nClasses)
134 
135  # Pass a testing data set and the trained model to the algorithm
136  algorithm.input.setTable(classifier.prediction.data, testData)
137  algorithm.input.setModel(classifier.prediction.model, trainingResult.get(classifier.training.model))
138 
139  # Predict Naive Bayes values (Result class from classifier.prediction)
140  predictionResult = algorithm.compute() # Retrieve the algorithm results
141 
142 
143 def printResults():
144  printNumericTables(
145  testGroundTruth, predictionResult.get(classifier.prediction.prediction),
146  "Ground truth", "Classification results",
147  "NaiveBayes classification results (first 20 observations):", 20, flt64=False
148  )
149 
150 if __name__ == "__main__":
151 
152  trainModel()
153  testModel()
154  printResults()

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