From 466bfd2836c4f5add545fe5f683cf0b7b9045304 Mon Sep 17 00:00:00 2001
From: Fabian Pedregosa <fabian.pedregosa@inria.fr>
Date: Thu, 29 Apr 2010 08:12:54 +0000
Subject: [PATCH] Add more tests to test_svm.py
git-svn-id: https://scikit-learn.svn.sourceforge.net/svnroot/scikit-learn/trunk@712 22fbfee3-77ab-4535-9bad-27d1bd3bc7d8
---
scikits/learn/tests/test_svm.py | 44 ++++++++++++++++++++++++++++-----
1 file changed, 38 insertions(+), 6 deletions(-)
diff --git a/scikits/learn/tests/test_svm.py b/scikits/learn/tests/test_svm.py
index 81e9331632..2ad5f323fc 100644
--- a/scikits/learn/tests/test_svm.py
+++ b/scikits/learn/tests/test_svm.py
@@ -25,7 +25,7 @@ def test_CSVC():
the second one with three classes. We check for predicted values
and estimated parameters.
- TODO: check with different parameters of C
+ TODO: check with different parameters of C, nonlinear kernel
"""
clf = svm.SVC(kernel='linear')
@@ -36,6 +36,7 @@ def test_CSVC():
assert_array_equal(clf.intercept_, [0.])
assert_array_equal(pred, true_result)
+ # the same with other dataset
clf.fit(X2, Y2)
pred = clf.predict(T2)
assert_array_almost_equal(clf.dual_coef_,
@@ -66,15 +67,42 @@ def test_SVR():
assert_array_almost_equal(clf.intercept_, [1.5])
assert_array_almost_equal(pred, [1.1, 2.3, 2.5])
+ # the same with kernel='rbf'
+ clf = svm.SVR(kernel='rbf')
+ clf.fit(X, Y)
+ pred = clf.predict(T)
+
+ assert_array_almost_equal(clf.dual_coef_,
+ [[-0.01441007, -0.51530606, -0.01365979,
+ 0.51569493, 0.01387495, 0.01380604]])
+ assert_raises(NotImplementedError, lambda: clf.coef_)
+ assert_array_almost_equal(clf.support_, X)
+ assert_array_almost_equal(clf.intercept_, [ 1.49997261])
+ assert_array_almost_equal(pred, [ 1.10001274, 1.86682485, 1.73300377])
+
def test_oneclass():
"""
FIXME: this does nothing
"""
- clf = svm.OneClassSVM()
+ clf = svm.OneClassSVM(kernel='linear')
+ clf.fit(X, Y)
+ pred = clf.predict(T)
+
+ assert_array_almost_equal(pred, [1, 1, 1])
+ assert_array_almost_equal(clf.intercept_, [0])
+ assert_array_almost_equal(clf.dual_coef_, [[.666, 1, .333, 1]], decimal=3)
+ assert_array_almost_equal(clf.coef_, [[0, 0]])
+
+ # the same with rbf kernel
+ clf = svm.OneClassSVM(kernel='rbf')
clf.fit(X, Y)
- assert_array_equal(Y, [1, 1, 1, 2, 2, 2])
+ pred = clf.predict(T)
+ assert_array_almost_equal(pred, [1, -1, -1])
+ assert_array_almost_equal(clf.intercept_, [-1.3514943])
+ assert_array_almost_equal(clf.dual_coef_, [[ 0.75061969, 0.74938031, 0.74996915, 0.75003085]])
+ assert_raises(NotImplementedError, lambda: clf.coef_)
def test_tweak_params():
@@ -98,7 +126,8 @@ def test_tweak_params():
def test_probability():
"""
- predict probabilities.
+ Predict probabilities using SVC
+
FIXME: is it harmless that we obtain slightly different results on
different operating systems ? (that is why we only check for 1
decimal precission)
@@ -131,6 +160,9 @@ def test_error():
assert_array_equal(clf.predict(T), [1, 2, 2])
def test_LinearSVC():
+ """
+ Test basic routines using LinearSVC
+ """
clf = svm.LinearSVC()
clf.fit(X, Y)
@@ -162,7 +194,7 @@ def test_coef_and_intercept_SVC_vs_LinearSVC():
linsvc.fit(X, Y)
assert_array_equal(linsvc.coef_.shape, svc.coef_.shape)
- assert_array_almost_equal(linsvc.coef_, svc.coef_, 5)
- assert_array_almost_equal(linsvc.intercept_, svc.intercept_, 6)
+ assert_array_almost_equal(linsvc.coef_, svc.coef_, decimal=5)
+ assert_array_almost_equal(linsvc.intercept_, svc.intercept_, decimal=5)
--
GitLab