def test_GMM_Unsup(genes, gene_len):
"""
Test that if unsupervised GMM method works
"""
k_min = 2
k_max = 6
num_class = 2
cov_type = 'full'
seed = 0
count_vect = CountVectorizer(analyzer='char', ngram_range=(k_min, k_max))
X = count_vect.fit_transform(genes)
chars = count_vect.get_feature_names()
kmers = X.toarray()
kmer_freq = []
for i in range(len(genes)):
kmer_freq.append(kmers[i] / gene_len[i])
kmer_table = pd.DataFrame(kmer_freq, columns=chars)
gmm = GMM(n_components=num_class,
covariance_type=cov_type,
random_state=seed).fit(kmer_table)
gmm.init_params = 'random'
labels = gmm.predict(kmer_table)
assert all(labels == [0, 0, 1, 0])
def get_predictions(path, k_min, k_max, num_class, cov_type, seed):
kmer_table = get_kmer_table(path, k_min, k_max)
gmm = GMM(n_components=num_class,
covariance_type=cov_type,
random_state=seed).fit(kmer_table)
gmm.init_params = 'random'
labels = gmm.predict(kmer_table)
return labels
def get_predictions_semi(path, k_min, k_max, num_class, cov_type, seed,
labels):
targets = []
kmer_table = get_kmer_table(path, k_min, k_max)
finalDf = pd.concat([kmer_table, pd.Series(labels)], axis=1)
gmm = GMM(n_components=num_class,
covariance_type=cov_type,
random_state=seed)
gmm.init_params = 'random'
for i in range(num_class):
if (i in list(finalDf.Labels)):
targets.append(i)
if (len(targets) == num_class):
#print("Yes")
gmm.means_init = np.array(
[kmer_table[finalDf.Labels == i].mean(axis=0) for i in targets])
gmm.fit(kmer_table)
predictions = gmm.predict(kmer_table)
return predictions
