def test_similarity(self):
cm = pd.DataFrame.from_dict(
{
"c1": [5, -1, 1, 2, 0],
"c2": [5, 4, 0, 2, 1],
"c3": [4, 4, 3, 1, 1],
"c4": [-1, 4, 0, 1, -1],
},
orient="index",
columns=["a", "b", "c", "d", "e"],
)
self.assertEqual(
dissimilarity_functions.hamming_similarity_without_missing(
list(cm.loc["c1", :]), list(cm.loc["c3", :]), -1),
0,
)
self.assertEqual(
dissimilarity_functions.hamming_similarity_without_missing(
list(cm.loc["c2", :]), list(cm.loc["c3", :]), -1),
2,
)
self.assertEqual(
dissimilarity_functions.hamming_similarity_without_missing(
list(cm.loc["c3", :]), list(cm.loc["c4", :]), -1),
2,
)
def test_hamming_similarity_without_missing_priors(self):
similarity = dissimilarity_functions.hamming_similarity_without_missing(
self.s1, self.s2, -1, weights=self.nlweights)
expected_similarity = np.sum(
[-np.log(self.priors[1][1]), -np.log(self.priors[4][1])])
self.assertEqual(similarity, expected_similarity)
def test_similarity_weighted(self):
cm = pd.DataFrame.from_dict(
{
"c1": [5, -1, 1, 2, 0],
"c2": [5, 4, 0, 2, 1],
"c3": [4, 4, 3, 1, 1],
"c4": [-1, 4, 0, 1, -1],
},
orient="index",
columns=["a", "b", "c", "d", "e"],
)
weights = {
0: {
5: 2
},
1: {
4: 2
},
2: {
1: 1,
3: 1
},
3: {
1: 1,
2: 1
},
4: {
1: 3
},
}
self.assertEqual(
dissimilarity_functions.hamming_similarity_without_missing(
list(cm.loc["c1", :]), list(cm.loc["c3", :]), -1, weights),
0,
)
self.assertEqual(
dissimilarity_functions.hamming_similarity_without_missing(
list(cm.loc["c2", :]), list(cm.loc["c3", :]), -1, weights),
5,
)
def test_hamming_similarity_without_missing_all_missing(self):
similarity = dissimilarity_functions.hamming_similarity_without_missing(
self.s1, self.all_missing, -1, weights=self.nlweights)
self.assertEqual(similarity, 0)
def test_hamming_similarity_without_missing_no_priors(self):
similarity = dissimilarity_functions.hamming_similarity_without_missing(
self.s1, self.s2, -1)
self.assertEqual(similarity, 2)
def test_hamming_similarity_without_missing_identical(self):
similarity = dissimilarity_functions.hamming_similarity_without_missing(
self.s1, self.s1, -1)
self.assertEqual(similarity, 3)