def predict(cls, experiment, sample): name = experiment.name exp_locator = Locator(experiment.id, experiment.train_data, experiment.test_data) if 'LR_' in name: lr_dataset_maker = LRPimaIndiansDatasetMaker(exp_locator) X = lr_dataset_maker.make_one_sample(sample) logistic_regression = DataLoader.load(exp_locator.get_model_file_path()) prediction = logistic_regression.predict(X) prediction = prediction.tolist() return prediction
def test(cls, experiment): name = experiment.name exp_locator = Locator(experiment.id, experiment.train_data, experiment.test_data) if 'LR_' in name: lr_dataset_maker = LRPimaIndiansDatasetMaker(exp_locator) X, y = lr_dataset_maker.make_test_dataset() logistic_regression = DataLoader.load(exp_locator.get_model_file_path()) test_result = logistic_regression.test(X, y) exp_result = json.loads(experiment.result) exp_result.append(test_result) experiment.result = json.dumps(exp_result) return experiment raise Exception("No valid Experiment Name to train Experiment")
def is_experiment_trained(cls, experiment): exp_locator = Locator(experiment.id, experiment.train_data, experiment.test_data) if os.path.exists(exp_locator.get_model_file_path()): return True else: return False