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

mn_naive_bayes_csr_online.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_CSR_ONLINE"></a>
42 ## \example mn_naive_bayes_csr_online.py
43 
44 import os
45 import sys
46 
47 from daal.algorithms.multinomial_naive_bayes import prediction, training
48 from daal.algorithms import classifier
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 printNumericTables, createSparseTable
55 
56 DAAL_PREFIX = os.path.join('..', 'data')
57 
58 # Input data set parameters
59 trainDatasetFileNames = [
60  os.path.join(DAAL_PREFIX, 'batch', 'naivebayes_train_csr.csv'),
61  os.path.join(DAAL_PREFIX, 'batch', 'naivebayes_train_csr.csv'),
62  os.path.join(DAAL_PREFIX, 'batch', 'naivebayes_train_csr.csv'),
63  os.path.join(DAAL_PREFIX, 'batch', 'naivebayes_train_csr.csv')
64 ]
65 
66 trainGroundTruthFileNames = [
67  os.path.join(DAAL_PREFIX, 'batch', 'naivebayes_train_labels.csv'),
68  os.path.join(DAAL_PREFIX, 'batch', 'naivebayes_train_labels.csv'),
69  os.path.join(DAAL_PREFIX, 'batch', 'naivebayes_train_labels.csv'),
70  os.path.join(DAAL_PREFIX, 'batch', 'naivebayes_train_labels.csv')
71 ]
72 
73 testDatasetFileName = os.path.join(DAAL_PREFIX, 'batch', 'naivebayes_test_csr.csv')
74 testGroundTruthFileName = os.path.join(DAAL_PREFIX, 'batch', 'naivebayes_test_labels.csv')
75 
76 nTrainVectorsInBlock = 8000
77 nTestObservations = 2000
78 nClasses = 20
79 nBlocks = 4
80 
81 trainingResult = None
82 predictionResult = None
83 trainData = [0] * nBlocks
84 testData = None
85 
86 
87 def trainModel():
88  global trainData, trainingResult
89 
90  # Create an algorithm object to train the Naive Bayes model
91  algorithm = training.Online(nClasses, method=training.fastCSR)
92 
93  for i in range(nBlocks):
94  # Read trainDatasetFileNames and create a numeric table to store the input data
95  trainData[i] = createSparseTable(trainDatasetFileNames[i])
96  trainLabelsSource = FileDataSource(
97  trainGroundTruthFileNames[i], DataSourceIface.doAllocateNumericTable,
98  DataSourceIface.doDictionaryFromContext
99  )
100 
101  trainLabelsSource.loadDataBlock(nTrainVectorsInBlock)
102 
103  # Pass a training data set and dependent values to the algorithm
104  algorithm.input.set(classifier.training.data, trainData[i])
105  algorithm.input.set(classifier.training.labels, trainLabelsSource.getNumericTable())
106 
107  # Build the Naive Bayes model
108  algorithm.compute()
109 
110  # Finalize the Naive Bayes model and retrieve the algorithm results
111  trainingResult = algorithm.finalizeCompute()
112 
113 
114 def testModel():
115  global predictionResult, testData
116 
117  # Read testDatasetFileName and create a numeric table to store the input data
118  testData = createSparseTable(testDatasetFileName)
119 
120  # Create an algorithm object to predict Naive Bayes values
121  algorithm = prediction.Batch(nClasses, method=prediction.fastCSR)
122 
123  # Pass a testing data set and the trained model to the algorithm
124  algorithm.input.setTable(classifier.prediction.data, testData)
125  algorithm.input.setModel(classifier.prediction.model, trainingResult.get(classifier.training.model))
126 
127  # Predict Naive Bayes values (Result class from classifier.prediction)
128  predictionResult = algorithm.compute() # Retrieve the algorithm results
129 
130 
131 def printResults():
132 
133  testGroundTruth = FileDataSource(
134  testGroundTruthFileName, DataSourceIface.doAllocateNumericTable,
135  DataSourceIface.doDictionaryFromContext
136  )
137  testGroundTruth.loadDataBlock(nTestObservations)
138 
139  printNumericTables(
140  testGroundTruth.getNumericTable(),
141  predictionResult.get(classifier.prediction.prediction),
142  "Ground truth", "Classification results",
143  "NaiveBayes classification results (first 20 observations):", 20, 15, flt64=False
144  )
145 
146 if __name__ == "__main__":
147 
148  trainModel()
149  testModel()
150  printResults()

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