Doctorado-ML/mufs
1
0
Fork 0
mirror of https://github.com/Doctorado-ML/mufs.git synced 2025-08-17 08:35:52 +00:00
Code Issues Packages Projects Releases Wiki Activity

Add max_features to selection

Browse Source 
Add first approach to continuous variables
This commit is contained in:
Ricardo Montañana Gómez
2021-06-01 11:30:52 +02:00
parent 39fbdf73a7
commit 794374fe8c
6 changed files with 714 additions and 158 deletions
Download Patch File Download Diff File Expand all files Collapse all files

43
mfs/tests/MFS_test.py
View File

@@ -9,11 +9,11 @@ class MFS_test(unittest.TestCase):
def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs)
mdlp = MDLP(random_state=1)
X, self.y_w = load_wine(return_X_y=True)
self.X_w = mdlp.fit_transform(X, self.y_w).astype("int64")
X, self.y_i = load_iris(return_X_y=True)
self.X_wc, self.y_w = load_wine(return_X_y=True)
self.X_w = mdlp.fit_transform(self.X_wc, self.y_w).astype("int64")
self.X_ic, self.y_i = load_iris(return_X_y=True)
mdlp = MDLP(random_state=1)
self.X_i = mdlp.fit_transform(X, self.y_i).astype("int64")
self.X_i = mdlp.fit_transform(self.X_ic, self.y_i).astype("int64")
def assertListAlmostEqual(self, list1, list2, tol=7):
self.assertEqual(len(list1), len(list2))
@@ -21,16 +21,16 @@ class MFS_test(unittest.TestCase):
self.assertAlmostEqual(a, b, tol)
def test_initialize(self):
mfs = MFS(max_features=100)
mfs = MFS()
mfs.fcbf(self.X_w, self.y_w, 0.05)
mfs._initialize()
mfs._initialize(self.X_w, self.y_w)
self.assertIsNone(mfs.get_results())
self.assertListEqual([], mfs.get_scores())
self.assertDictEqual({}, mfs._su_features)
self.assertIsNone(mfs._su_labels)
def test_csf_wine(self):
mfs = MFS(max_features=100)
mfs = MFS()
expected = [6, 12, 9, 4, 10, 0]
self.assertListAlmostEqual(
expected, mfs.cfs(self.X_w, self.y_w).get_results()
@@ -45,6 +45,23 @@ class MFS_test(unittest.TestCase):
]
self.assertListAlmostEqual(expected, mfs.get_scores())
def test_csf_wine_cont(self):
mfs = MFS(discrete=False)
expected = [6, 11, 9, 0, 12, 5]
self.assertListAlmostEqual(
expected, mfs.cfs(self.X_wc, self.y_w).get_results()
)
expected = [
0.5218299405215557,
0.602513857132804,
0.4877384978817362,
0.3743688234383051,
0.28795671854246285,
0.2309165735173175,
]
# self.assertListAlmostEqual(expected, mfs.get_scores())
print(expected, mfs.get_scores())
def test_csf_max_features(self):
mfs = MFS(max_features=3)
expected = [6, 12, 9]
@@ -59,7 +76,7 @@ class MFS_test(unittest.TestCase):
self.assertListAlmostEqual(expected, mfs.get_scores())
def test_csf_iris(self):
mfs = MFS(max_features=100)
mfs = MFS()
expected = [3, 2, 0, 1]
computed = mfs.cfs(self.X_i, self.y_i).get_results()
self.assertListAlmostEqual(expected, computed)
@@ -72,7 +89,7 @@ class MFS_test(unittest.TestCase):
self.assertListAlmostEqual(expected, mfs.get_scores())
def test_fcbf_wine(self):
mfs = MFS(max_features=100)
mfs = MFS()
computed = mfs.fcbf(self.X_w, self.y_w, threshold=0.05).get_results()
expected = [6, 9, 12, 0, 11, 4]
self.assertListAlmostEqual(expected, computed)
@@ -99,7 +116,7 @@ class MFS_test(unittest.TestCase):
self.assertListAlmostEqual(expected, mfs.get_scores())
def test_fcbf_iris(self):
mfs = MFS(max_features=100)
mfs = MFS()
computed = mfs.fcbf(self.X_i, self.y_i, threshold=0.05).get_results()
expected = [3, 2]
self.assertListAlmostEqual(expected, computed)
@@ -107,7 +124,7 @@ class MFS_test(unittest.TestCase):
self.assertListAlmostEqual(expected, mfs.get_scores())
def test_compute_su_labels(self):
mfs = MFS(max_features=100)
mfs = MFS()
mfs.fcbf(self.X_i, self.y_i, threshold=0.05)
expected = [0.0, 0.0, 0.810724587460511, 0.870521418179061]
self.assertListAlmostEqual(expected, mfs._compute_su_labels().tolist())
@@ -115,12 +132,12 @@ class MFS_test(unittest.TestCase):
self.assertListAlmostEqual([1, 2, 3, 4], mfs._compute_su_labels())
def test_invalid_threshold(self):
mfs = MFS(max_features=100)
mfs = MFS()
with self.assertRaises(ValueError):
mfs.fcbf(self.X_i, self.y_i, threshold=1e-15)
def test_fcbf_exit_threshold(self):
mfs = MFS(max_features=100)
mfs = MFS()
computed = mfs.fcbf(self.X_w, self.y_w, threshold=0.4).get_results()
expected = [6, 9, 12]
self.assertListAlmostEqual(expected, computed)

77
mfs/tests/Metrics_test.py
View File

@@ -1,5 +1,6 @@
import unittest
from sklearn.datasets import load_iris
import numpy as np
from sklearn.datasets import load_iris, load_wine
from mdlp import MDLP
from ..Selection import Metrics
@@ -8,12 +9,10 @@ class Metrics_test(unittest.TestCase):
def __init__(self, *args, **kwargs):
super().__init__(*args, **kwargs)
mdlp = MDLP(random_state=1)
X, self.y = load_iris(return_X_y=True)
self.X = mdlp.fit_transform(X, self.y).astype("int64")
self.m, self.n = self.X.shape
# @classmethod
# def setup(cls):
self.X_i_c, self.y_i = load_iris(return_X_y=True)
self.X_i = mdlp.fit_transform(self.X_i_c, self.y_i).astype("int64")
self.X_w_c, self.y_w = load_wine(return_X_y=True)
self.X_w = mdlp.fit_transform(self.X_w_c, self.y_w).astype("int64")
def test_entropy(self):
metric = Metrics()
@@ -24,12 +23,51 @@ class Metrics_test(unittest.TestCase):
([1, 1, 1, 5, 2, 2, 3, 3, 3], 4, 0.9455305560363263),
([1, 1, 1, 2, 2, 3, 3, 3, 5], 4, 0.9455305560363263),
([1, 1, 5], 2, 0.9182958340544896),
(self.y, 3, 0.999999999),
(self.y_i, 3, 0.999999999),
]
for dataset, base, entropy in datasets:
computed = metric.entropy(dataset, base)
self.assertAlmostEqual(entropy, computed)
def test_differential_entropy(self):
metric = Metrics()
datasets = [
([0, 0, 0, 0, 1, 1, 1, 1], 6, 1.0026709900837547096),
([0, 1, 0, 2, 1, 2], 5, 1.3552453009332424),
([0, 0, 0, 0, 0, 0, 0, 2, 2, 2], 7, 1.7652626150881443),
([1, 1, 1, 5, 2, 2, 3, 3, 3], 8, 1.9094631320594582),
([1, 1, 1, 2, 2, 3, 3, 3, 5], 8, 1.9094631320594582),
([1, 1, 5], 2, 2.5794415416798357),
(self.X_i_c, 37, 3.06627326925228),
(self.X_w_c, 37, 63.13827518897429),
]
for dataset, base, entropy in datasets:
computed = metric.differential_entropy(
np.array(dataset, dtype="float64"), base
)
self.assertAlmostEqual(entropy, computed, msg=str(dataset))
expected = [
1.6378708764142766,
2.0291571802275037,
0.8273865123744271,
3.203935772642847,
4.859193341386733,
1.3707315434976266,
1.8794952925706312,
-0.2983180654207054,
1.4521478934625076,
2.834404839362728,
0.4894081282811191,
1.361210381692561,
7.6373991502818175,
]
n_samples, n_features = self.X_w_c.shape
for c, res_expected in zip(range(n_features), expected):
computed = metric.differential_entropy(
self.X_w_c[:, c], n_samples - 1
)
self.assertAlmostEqual(computed, res_expected)
def test_conditional_entropy(self):
metric = Metrics()
results_expected = [
@@ -39,7 +77,7 @@ class Metrics_test(unittest.TestCase):
0.13032469395094992,
]
for expected, col in zip(results_expected, range(self.n)):
computed = metric.conditional_entropy(self.X[:, col], self.y, 3)
computed = metric.conditional_entropy(self.X_i[:, col], self.y, 3)
self.assertAlmostEqual(expected, computed)
self.assertAlmostEqual(
0.6309297535714573,
@@ -62,7 +100,7 @@ class Metrics_test(unittest.TestCase):
0.8696753060490499,
]
for expected, col in zip(results_expected, range(self.n)):
computed = metric.information_gain(self.X[:, col], self.y, 3)
computed = metric.information_gain(self.X_i[:, col], self.y, 3)
self.assertAlmostEqual(expected, computed)
# https://planetcalc.com/8419/
# ?_d=FrDfFN2COAhqh9Pb5ycqy5CeKgIOxlfSjKgyyIR.Q5L0np-g-hw6yv8M1Q8_
@@ -73,7 +111,7 @@ class Metrics_test(unittest.TestCase):
1.378402748,
]
for expected, col in zip(results_expected, range(self.n)):
computed = metric.information_gain(self.X[:, col], self.y, 2)
computed = metric.information_gain(self.X_i[:, col], self.y, 2)
self.assertAlmostEqual(expected, computed)
def test_symmetrical_uncertainty(self):
@@ -85,5 +123,20 @@ class Metrics_test(unittest.TestCase):
0.870521418179061,
]
for expected, col in zip(results_expected, range(self.n)):
computed = metric.symmetrical_uncertainty(self.X[:, col], self.y)
computed = metric.symmetrical_uncertainty(self.X_i[:, col], self.y)
self.assertAlmostEqual(expected, computed)
def test_symmetrical_uncertainty_continuous(self):
metric = Metrics()
results_expected = [
0.33296547388990266,
0.19068147573570668,
0.810724587460511,
0.870521418179061,
]
for expected, col in zip(results_expected, range(self.n)):
computed = metric.symmetrical_unc_continuous(
self.X_i[:, col], self.y
)
print(computed)
# self.assertAlmostEqual(expected, computed)