| 123456789101112131415161718192021222324252627282930313233343536373839404142434445464748495051525354555657585960616263646566676869707172737475767778798081828384858687888990919293 |
- import unittest
- import numpy as np
- from federatedml.util import consts
- from federatedml.evaluation.metrics import classification_metric, clustering_metric, regression_metric
- from federatedml.evaluation.metric_interface import MetricInterface
- class TestEvaluation(unittest.TestCase):
- def setUp(self):
- self.bin_score = np.random.random(100)
- self.bin_label = (self.bin_score > 0.5) + 0
- self.reg_score = np.random.random(100) * 10
- self.reg_label = np.random.random(100) * 10
- self.multi_score = np.random.randint([4 for i in range(50)])
- self.multi_label = np.random.randint([4 for i in range(50)])
- self.clustering_score = np.random.randint([4 for i in range(50)])
- self.clustering_label = np.random.randint([3 for i in range(50)])
- self.psi_train_score = np.random.random(10000)
- self.psi_train_label = (self.psi_train_score > 0.5) + 0
- self.psi_val_score = np.random.random(1000)
- self.psi_val_label = (self.psi_val_score > 0.5) + 0
- def test_regression(self):
- print('testing regression metric')
- regression_metric.R2Score().compute(self.reg_score, self.reg_label)
- regression_metric.MSE().compute(self.reg_score, self.reg_label)
- regression_metric.RMSE().compute(self.reg_score, self.reg_label)
- regression_metric.ExplainedVariance().compute(self.reg_score, self.reg_label)
- regression_metric.Describe().compute(self.reg_score)
- def test_binary(self):
- print('testing binary')
- interface = MetricInterface(pos_label=1, eval_type=consts.BINARY)
- interface.auc(self.bin_label, self.bin_score)
- interface.confusion_mat(self.bin_label, self.bin_score)
- interface.ks(self.bin_label, self.bin_score)
- interface.accuracy(self.bin_label, self.bin_score)
- interface.f1_score(self.bin_label, self.bin_score)
- interface.gain(self.bin_label, self.bin_score)
- interface.lift(self.bin_label, self.bin_score)
- interface.quantile_pr(self.bin_label, self.bin_score)
- interface.precision(self.bin_label, self.bin_score)
- interface.recall(self.bin_label, self.bin_score)
- interface.roc(self.bin_label, self.bin_score)
- def test_psi(self):
- interface = MetricInterface(pos_label=1, eval_type=consts.BINARY)
- interface.psi(
- self.psi_train_score,
- self.psi_val_score,
- train_labels=self.psi_train_label,
- validate_labels=self.psi_val_label)
- def test_multi(self):
- print('testing multi')
- interface = MetricInterface(eval_type=consts.MULTY, pos_label=1)
- interface.precision(self.multi_label, self.multi_score)
- interface.recall(self.multi_label, self.multi_score)
- interface.accuracy(self.multi_label, self.multi_score)
- def test_clustering(self):
- print('testing clustering')
- interface = MetricInterface(eval_type=consts.CLUSTERING, pos_label=1)
- interface.confusion_mat(self.clustering_label, self.clustering_score)
- def test_newly_added(self):
- print('testing newly added')
- binary_data = list(
- zip([i for i in range(len(self.psi_train_score))], self.psi_train_score))
- classification_metric.Distribution().compute(binary_data, binary_data)
- multi_data = list(
- zip([i for i in range(len(self.multi_score))], self.multi_score))
- classification_metric.Distribution().compute(multi_data, multi_data)
- classification_metric.KSTest().compute(self.multi_score, self.multi_score)
- classification_metric.KSTest().compute(
- self.psi_train_score, self.psi_val_score)
- classification_metric.AveragePrecisionScore().compute(
- self.psi_train_score,
- self.psi_val_score,
- self.psi_train_label,
- self.psi_val_label)
- if __name__ == '__main__':
- unittest.main()
|
