def test_getNumpy_type(self):
X, y = datasets.RiverFlow1().get_numpy()
self.assertTrue(type(X) is numpy.ndarray)
self.assertTrue(X.dtype is numpy.dtype('float32'))
self.assertTrue(type(y) is numpy.ndarray)
self.assertTrue(y.dtype is numpy.dtype('float32'))
import amorf.datasets as ds
import amorf.problemTransformation as pt
import amorf.metrics as metrics
import numpy as np
from sklearn.model_selection import KFold
edm = ds.EDM().get_numpy()
rf1 = ds.RiverFlow1().get_numpy()
wq = ds.WaterQuality().get_numpy()
transCond = ds.TransparentConductors().get_numpy()
dataset_names = ['EDM', 'RF1', 'Water Quality', 'Transparent Conductors']
datasets = [edm, rf1, wq, transCond]
results_datasets = []
for dataset in datasets:
selectors = [
'linear', 'kneighbors', 'adaboost', 'gradientboost', 'mlp', 'svr',
'xgb'
]
all_results = []
for selector in selectors:
SM = pt.SingleTargetMethod(selector)
X = dataset[0]
y = dataset[1]
kf = KFold(n_splits=5, random_state=1, shuffle=True)
selector_results = []
for train_index, test_index in kf.split(X):
prediction = SM.fit(X[train_index],
y[train_index]).predict(X[test_index])
result = metrics.average_relative_root_mean_squared_error(
prediction, y[test_index])
def test_getNumpy_dimensions(self):
X, y = datasets.RiverFlow1().get_numpy()
self.assertEqual(len(X), len(y))
self.assertEqual(len(X[0, :]), 64)
self.assertEqual(len(y[0, :]), 8)