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

logitboost_dense_batch.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 
42 
43 
44 import os
45 import sys
46 
47 from daal.algorithms.logitboost import prediction, training
48 from daal.algorithms import classifier
49 from daal.data_management import (
50  FileDataSource, DataSourceIface, NumericTableIface, HomogenNumericTable, MergedNumericTable
51 )
52 
53 utils_folder = os.path.realpath(os.path.abspath(os.path.dirname(os.path.dirname(__file__))))
54 if utils_folder not in sys.path:
55  sys.path.insert(0, utils_folder)
56 from utils import printNumericTables
57 
58 DAAL_PREFIX = os.path.join('..', 'data')
59 
60 # Input data set parameters
61 trainDatasetFileName = os.path.join(DAAL_PREFIX, 'batch', 'logitboost_train.csv')
62 testDatasetFileName = os.path.join(DAAL_PREFIX, 'batch', 'logitboost_test.csv')
63 
64 nFeatures = 20
65 nClasses = 5
66 
67 # LogitBoost algorithm parameters
68 maxIterations = 100 # Maximum number of terms in additive regression
69 accuracyThreshold = 0.01 # Training accuracy
70 
71 # Model object for the LogitBoost algorithm
72 model = None
73 predictionResult = None
74 testGroundTruth = None
75 
76 
77 def trainModel():
78  global model
79 
80  # Initialize FileDataSource<CSVFeatureManager> to retrieve the input data from a .csv file
81  trainDataSource = FileDataSource(
82  trainDatasetFileName,
83  DataSourceIface.notAllocateNumericTable,
84  DataSourceIface.doDictionaryFromContext
85  )
86 
87  # Create Numeric Tables for training data and labels
88  trainData = HomogenNumericTable(nFeatures, 0, NumericTableIface.doNotAllocate)
89  trainGroundTruth = HomogenNumericTable(1, 0, NumericTableIface.doNotAllocate)
90  mergedData = MergedNumericTable(trainData, trainGroundTruth)
91 
92  # Retrieve the data from the input file
93  trainDataSource.loadDataBlock(mergedData)
94 
95  # Create an algorithm object to train the LogitBoost model
96  algorithm = training.Batch(nClasses)
97  algorithm.parameter.maxIterations = maxIterations
98  algorithm.parameter.accuracyThreshold = accuracyThreshold
99 
100  # Pass the training data set and dependent values to the algorithm
101  algorithm.input.set(classifier.training.data, trainData)
102  algorithm.input.set(classifier.training.labels, trainGroundTruth)
103 
104  # Train the LogitBoost model and retrieve the results of the training algorithm
105  trainingResult = algorithm.compute()
106  model = trainingResult.get(classifier.training.model)
107 
108 
109 def testModel():
110  global testGroundTruth, predictionResult
111 
112  # Initialize FileDataSource<CSVFeatureManager> to retrieve the test data from a .csv file
113  testDataSource = FileDataSource(
114  testDatasetFileName,
115  DataSourceIface.notAllocateNumericTable,
116  DataSourceIface.doDictionaryFromContext
117  )
118 
119  # Create Numeric Tables for testing data and labels
120  testData = HomogenNumericTable(nFeatures, 0, NumericTableIface.doNotAllocate)
121  testGroundTruth = HomogenNumericTable(1, 0, NumericTableIface.doNotAllocate)
122  mergedData = MergedNumericTable(testData, testGroundTruth)
123 
124  # Retrieve the data from input file
125  testDataSource.loadDataBlock(mergedData)
126 
127  # Create algorithm objects for LogitBoost prediction with the default method
128  algorithm = prediction.Batch(nClasses)
129 
130  # Pass the testing data set and trained model to the algorithm
131  algorithm.input.setTable(classifier.prediction.data, testData)
132  algorithm.input.setModel(classifier.prediction.model, model)
133 
134  # Compute prediction results and retrieve algorithm results
135  # (Result class from classifier.prediction)
136  predictionResult = algorithm.compute()
137 
138 
139 def printResults():
140 
141  printNumericTables(
142  testGroundTruth,
143  predictionResult.get(classifier.prediction.prediction),
144  "Ground truth", "Classification results",
145  "LogitBoost classification results (first 20 observations):", 20, flt64=False
146  )
147 
148 if __name__ == "__main__":
149 
150  trainModel()
151  testModel()
152  printResults()

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