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

lin_reg_qr_dense_batch.py

1 # file: lin_reg_qr_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 
43 
44 
45 import os
46 import sys
47 
48 from daal.algorithms.linear_regression import training, prediction
49 from daal.data_management import (
50  DataSourceIface, FileDataSource, HomogenNumericTable, MergedNumericTable, NumericTableIface
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 printNumericTable
57 
58 DAAL_PREFIX = os.path.join('..', 'data')
59 
60 # Input data set parameters
61 trainDatasetFileName = os.path.join(DAAL_PREFIX, 'batch', 'linear_regression_train.csv')
62 testDatasetFileName = os.path.join(DAAL_PREFIX, 'batch', 'linear_regression_test.csv')
63 
64 nFeatures = 10 # Number of features in training and testing data sets
65 nDependentVariables = 2 # Number of dependent variables that correspond to each observation
66 
67 trainingResult = None
68 predictionResult = None
69 
70 
71 def trainModel():
72  global trainingResult
73 
74  # Initialize FileDataSource<CSVFeatureManager> to retrieve the input data from a .csv file
75  trainDataSource = FileDataSource(
76  trainDatasetFileName, DataSourceIface.notAllocateNumericTable,
77  DataSourceIface.doDictionaryFromContext
78  )
79 
80  # Create Numeric Tables for training data and dependent variables
81  trainData = HomogenNumericTable(nFeatures, 0, NumericTableIface.doNotAllocate)
82  trainDependentVariables = HomogenNumericTable(
83  nDependentVariables, 0, NumericTableIface.doNotAllocate
84  )
85  mergedData = MergedNumericTable(trainData, trainDependentVariables)
86 
87  # Retrieve the data from input file
88  trainDataSource.loadDataBlock(mergedData)
89 
90  # Create an algorithm object to train the multiple linear regression model with a QR decomposition-based method
91  algorithm = training.Batch(method=training.qrDense)
92 
93  # Pass a training data set and dependent values to the algorithm
94  algorithm.input.set(training.data, trainData)
95  algorithm.input.set(training.dependentVariables, trainDependentVariables)
96 
97  # Build the multiple linear regression model and retrieve the algorithm results
98  trainingResult = algorithm.compute()
99  printNumericTable(trainingResult.get(training.model).getBeta(), "Linear Regression coefficients:")
100 
101 
102 def testModel():
103  global predictionResult
104 
105  # Initialize FileDataSource<CSVFeatureManager> to retrieve the test data from a .csv file
106  testDataSource = FileDataSource(
107  testDatasetFileName, DataSourceIface.doAllocateNumericTable,
108  DataSourceIface.doDictionaryFromContext
109  )
110 
111  # Create Numeric Tables for testing data and ground truth values
112  testData = HomogenNumericTable(nFeatures, 0, NumericTableIface.doNotAllocate)
113  testGroundTruth = HomogenNumericTable(nDependentVariables, 0, NumericTableIface.doNotAllocate)
114  mergedData = MergedNumericTable(testData, testGroundTruth)
115 
116  testDataSource.loadDataBlock(mergedData)
117 
118  # Create an algorithm object to predict values of multiple linear regression
119  algorithm = prediction.Batch()
120 
121  # Pass a testing data set and the trained model to the algorithm
122  algorithm.input.setTable(prediction.data, testData)
123  algorithm.input.setModel(prediction.model, trainingResult.get(training.model))
124 
125  # Predict values of multiple linear regression and retrieve the algorithm results
126  predictionResult = algorithm.compute()
127  printNumericTable(predictionResult.get(prediction.prediction), "Linear Regression prediction results: (first 10 rows):", 10)
128  printNumericTable(testGroundTruth, "Ground truth (first 10 rows):", 10)
129 
130 if __name__ == "__main__":
131 
132  trainModel()
133  testModel()

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