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

adaboost_dense_batch.py

1 # file: adaboost_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 from daal.algorithms.adaboost import prediction, training
49 from daal.algorithms import classifier
50 from daal.data_management import (
51  FileDataSource, DataSourceIface, HomogenNumericTable, MergedNumericTable, NumericTableIface
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 trainDatasetFileName = os.path.join(DAAL_PREFIX, 'batch', 'adaboost_train.csv')
63 
64 testDatasetFileName = os.path.join(DAAL_PREFIX, 'batch', 'adaboost_test.csv')
65 
66 nFeatures = 20
67 
68 trainingResult = None
69 predictionResult = None
70 testGroundTruth = None
71 
72 
73 def trainModel():
74  global trainingResult
75 
76  # Initialize FileDataSource<CSVFeatureManager> to retrieve the input data from a .csv file
77  trainDataSource = FileDataSource(
78  trainDatasetFileName, DataSourceIface.notAllocateNumericTable,
79  DataSourceIface.doDictionaryFromContext
80  )
81 
82  # Create Numeric Tables for training data and labels
83  trainData = HomogenNumericTable(nFeatures, 0, NumericTableIface.doNotAllocate)
84  trainGroundTruth = HomogenNumericTable(1, 0, NumericTableIface.doNotAllocate)
85  mergedData = MergedNumericTable(trainData, trainGroundTruth)
86 
87  # Retrieve the data from the input file
88  trainDataSource.loadDataBlock(mergedData)
89 
90  # Create an algorithm object to train the AdaBoost model
91  algorithm = training.Batch()
92 
93  # Pass the training data set and dependent values to the algorithm
94  algorithm.input.set(classifier.training.data, trainData)
95  algorithm.input.set(classifier.training.labels, trainGroundTruth)
96 
97  # Train the AdaBoost model and retrieve the results of the training algorithm
98  trainingResult = algorithm.compute()
99 
100 
101 def testModel():
102  global predictionResult, testGroundTruth
103 
104  # Initialize FileDataSource<CSVFeatureManager> to retrieve the test data from a .csv file
105  testDataSource = FileDataSource(
106  testDatasetFileName, DataSourceIface.notAllocateNumericTable,
107  DataSourceIface.doDictionaryFromContext
108  )
109 
110  # Create Numeric Tables for testing data and labels
111  testData = HomogenNumericTable(nFeatures, 0, NumericTableIface.doNotAllocate)
112  testGroundTruth = HomogenNumericTable(1, 0, NumericTableIface.doNotAllocate)
113  mergedData = MergedNumericTable(testData, testGroundTruth)
114 
115  # Retrieve the data from input file
116  testDataSource.loadDataBlock(mergedData)
117 
118  # Create algorithm objects for AdaBoost prediction with the default method
119  algorithm = prediction.Batch()
120 
121  # Pass the testing data set and trained model to the algorithm
122  algorithm.input.setTable(classifier.prediction.data, testData)
123  algorithm.input.setModel(classifier.prediction.model, trainingResult.get(classifier.training.model))
124 
125  # Compute prediction results and retrieve algorithm results
126  # (Result class from classifier.prediction)
127  predictionResult = algorithm.compute()
128 
129 def printResults():
130  printNumericTables(
131  testGroundTruth, predictionResult.get(classifier.prediction.prediction),
132  "Ground truth", "Classification results",
133  "AdaBoost classification results (first 20 observations):", 20, flt64=False
134  )
135 
136 if __name__ == "__main__":
137 
138  trainModel()
139  testModel()
140  printResults()

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