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

mn_naive_bayes_csr_online.py

1 # file: mn_naive_bayes_csr_online.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 
43 
44 
45 import os
46 import sys
47 
48 from daal.algorithms.multinomial_naive_bayes import prediction, training
49 from daal.algorithms import classifier
50 from daal.data_management import FileDataSource, DataSourceIface
51 
52 utils_folder = os.path.realpath(os.path.abspath(os.path.dirname(os.path.dirname(__file__))))
53 if utils_folder not in sys.path:
54  sys.path.insert(0, utils_folder)
55 from utils import printNumericTables, createSparseTable
56 
57 DAAL_PREFIX = os.path.join('..', 'data')
58 
59 # Input data set parameters
60 trainDatasetFileNames = [
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  os.path.join(DAAL_PREFIX, 'batch', 'naivebayes_train_csr.csv')
65 ]
66 
67 trainGroundTruthFileNames = [
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  os.path.join(DAAL_PREFIX, 'batch', 'naivebayes_train_labels.csv')
72 ]
73 
74 testDatasetFileName = os.path.join(DAAL_PREFIX, 'batch', 'naivebayes_test_csr.csv')
75 testGroundTruthFileName = os.path.join(DAAL_PREFIX, 'batch', 'naivebayes_test_labels.csv')
76 
77 nTrainVectorsInBlock = 8000
78 nTestObservations = 2000
79 nClasses = 20
80 nBlocks = 4
81 
82 trainingResult = None
83 predictionResult = None
84 trainData = [0] * nBlocks
85 testData = None
86 
87 
88 def trainModel():
89  global trainData, trainingResult
90 
91  # Create an algorithm object to train the Naive Bayes model
92  algorithm = training.Online(nClasses, method=training.fastCSR)
93 
94  for i in range(nBlocks):
95  # Read trainDatasetFileNames and create a numeric table to store the input data
96  trainData[i] = createSparseTable(trainDatasetFileNames[i])
97  trainLabelsSource = FileDataSource(
98  trainGroundTruthFileNames[i], DataSourceIface.doAllocateNumericTable,
99  DataSourceIface.doDictionaryFromContext
100  )
101 
102  trainLabelsSource.loadDataBlock(nTrainVectorsInBlock)
103 
104  # Pass a training data set and dependent values to the algorithm
105  algorithm.input.set(classifier.training.data, trainData[i])
106  algorithm.input.set(classifier.training.labels, trainLabelsSource.getNumericTable())
107 
108  # Build the Naive Bayes model
109  algorithm.compute()
110 
111  # Finalize the Naive Bayes model and retrieve the algorithm results
112  trainingResult = algorithm.finalizeCompute()
113 
114 
115 def testModel():
116  global predictionResult, testData
117 
118  # Read testDatasetFileName and create a numeric table to store the input data
119  testData = createSparseTable(testDatasetFileName)
120 
121  # Create an algorithm object to predict Naive Bayes values
122  algorithm = prediction.Batch(nClasses, method=prediction.fastCSR)
123 
124  # Pass a testing data set and the trained model to the algorithm
125  algorithm.input.setTable(classifier.prediction.data, testData)
126  algorithm.input.setModel(classifier.prediction.model, trainingResult.get(classifier.training.model))
127 
128  # Predict Naive Bayes values (Result class from classifier.prediction)
129  predictionResult = algorithm.compute() # Retrieve the algorithm results
130 
131 
132 def printResults():
133 
134  testGroundTruth = FileDataSource(
135  testGroundTruthFileName, DataSourceIface.doAllocateNumericTable,
136  DataSourceIface.doDictionaryFromContext
137  )
138  testGroundTruth.loadDataBlock(nTestObservations)
139 
140  printNumericTables(
141  testGroundTruth.getNumericTable(),
142  predictionResult.get(classifier.prediction.prediction),
143  "Ground truth", "Classification results",
144  "NaiveBayes classification results (first 20 observations):", 20, 15, flt64=False
145  )
146 
147 if __name__ == "__main__":
148 
149  trainModel()
150  testModel()
151  printResults()

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