22 from daal.algorithms
import gbt
23 from daal.algorithms.gbt.regression
import prediction, training
24 from daal.data_management
import (
25 FileDataSource, DataSourceIface, NumericTableIface,
26 HomogenNumericTable, MergedNumericTable, features
29 utils_folder = os.path.realpath(os.path.abspath(os.path.dirname(os.path.dirname(__file__))))
30 if utils_folder
not in sys.path:
31 sys.path.insert(0, utils_folder)
32 from utils
import printNumericTable
34 DAAL_PREFIX = os.path.join(
'..',
'data')
37 trainDatasetFileName = os.path.join(DAAL_PREFIX,
'batch',
'df_regression_train.csv')
38 testDatasetFileName = os.path.join(DAAL_PREFIX,
'batch',
'df_regression_test.csv')
47 predictionResult =
None
48 testGroundTruth =
None
55 trainDataSource = FileDataSource(
57 DataSourceIface.notAllocateNumericTable,
58 DataSourceIface.doDictionaryFromContext
62 trainData = HomogenNumericTable(nFeatures, 0, NumericTableIface.notAllocate)
63 trainGroundTruth = HomogenNumericTable(1, 0, NumericTableIface.notAllocate)
64 mergedData = MergedNumericTable(trainData, trainGroundTruth)
67 trainDataSource.loadDataBlock(mergedData)
70 dict = trainData.getDictionary()
73 dict[3].featureType = features.DAAL_CATEGORICAL
76 algorithm = training.Batch()
77 algorithm.parameter().maxIterations = maxIterations
80 algorithm.input.set(training.data, trainData)
81 algorithm.input.set(training.dependentVariable, trainGroundTruth)
84 trainingResult = algorithm.compute()
85 model = trainingResult.get(training.model)
88 global testGroundTruth, predictionResult
91 testDataSource = FileDataSource(
93 DataSourceIface.notAllocateNumericTable,
94 DataSourceIface.doDictionaryFromContext
98 testData = HomogenNumericTable(nFeatures, 0, NumericTableIface.notAllocate)
99 testGroundTruth = HomogenNumericTable(1, 0, NumericTableIface.notAllocate)
100 mergedData = MergedNumericTable(testData, testGroundTruth)
103 testDataSource.loadDataBlock(mergedData)
106 dict = testData.getDictionary()
109 dict[3].featureType = features.DAAL_CATEGORICAL
112 algorithm = prediction.Batch()
115 algorithm.input.setTable(prediction.data, testData)
116 algorithm.input.set(prediction.model, model)
119 predictionResult = algorithm.compute()
125 predictionResult.get(prediction.prediction),
126 "Gradient boosted trees prediction results (first 10 rows):", 10
130 "Ground truth (first 10 rows):", 10
133 if __name__ ==
"__main__":