def test_pass_pca_corr_pca_out():
X, y = iris_data()
pca = PCA(n_components=2)
X_pca = pca.fit_transform(X)
eigen = pca.explained_variance_
plot_pca_correlation_graph(X,
variables_names=['1', '2', '3', '4'],
X_pca=X_pca,
explained_variance=eigen)
def test_no_X_PCA_but_explained_variance():
with pytest.raises(ValueError,
match='If `explained variance` is not None, the '
'`X_pca` values should not be `None`.'):
X, y = iris_data()
pca = PCA(n_components=2)
pca.fit(X)
eigen = pca.explained_variance_
plot_pca_correlation_graph(X,
variables_names=['1', '2', '3', '4'],
X_pca=None,
explained_variance=eigen)
def test_X_PCA_but_no_explained_variance():
with pytest.raises(
ValueError,
match='If `X_pca` is not None, the `explained variance` '
'values should not be `None`.'):
X, y = iris_data()
pca = PCA(n_components=2)
X_pca = pca.fit_transform(X)
plot_pca_correlation_graph(X,
variables_names=['1', '2', '3', '4'],
X_pca=X_pca,
explained_variance=None)
def test_not_enough_components():
s = (
'Number of principal components must match the number of eigenvalues. Got 2 != 1'
)
with pytest.raises(ValueError, match=s):
X, y = iris_data()
pca = PCA(n_components=2)
X_pca = pca.fit_transform(X)
eigen = pca.explained_variance_
plot_pca_correlation_graph(X,
variables_names=['1', '2', '3', '4'],
X_pca=X_pca,
explained_variance=eigen[:-1])
def pca(x, dataset, mod):
"""
:param x:
:param dataset:
:return:
"""
xn = x[list(x.columns[np.argwhere(np.array(mod) == 'numeric')])]
xm = x[list(x.columns[np.argwhere(np.array(mod) == 'modal')])]
if dataset == 'kidney-disease':
n = 10
pca_instance = PCA(n_components=n)
pca_instance.fit(xn[list(xn.columns)])
xp = pca_instance.transform(xn)
var = sum(pca_instance.explained_variance_[:n]) * 100 / sum(
pca_instance.explained_variance_)
print(
'The {} principal components are responsible for {}% of the variance'
.format(n, var))
feature_names = list(xn.columns)
figure, correlation_matrix = plot_pca_correlation_graph(
xn, feature_names, figure_axis_size=10)
plt.savefig('Output/Images' + "/" +
'PCA_for_dataset_{}'.format(dataset))
plt.close()
xp = pd.DataFrame(data=xp, index=xn.index)
xp = pd.concat([xp, xm], axis=1)
return xp
elif dataset == 'bank-note':
n = 2
pca_instance = PCA(n_components=n)
pca_instance.fit(x)
xp = pca_instance.transform(x)
var = sum(pca_instance.explained_variance_[:n]) * 100 / sum(
pca_instance.explained_variance_)
print(
'The {} principal components are responsible for {}% of the variance'
.format(n, var))
feature_names = ['variance', 'skewness', 'curtosis', 'entropy']
figure, correlation_matrix = plot_pca_correlation_graph(
x, feature_names, figure_axis_size=10)
plt.savefig('Output/Images' + "/" +
'PCA_for_dataset_{}'.format(dataset))
plt.close()
return xp
g = sns.lmplot('PC1',
'PC2',
hue='Case of flush',
data=ml_data,
fit_reg=False,
scatter=True)
plt.show()
#Correlation Circle
features_name = [
'Blue 1-1', 'Blue 1-2', 'Green 1-1', 'Green 1-2', 'Red 1-1', 'Red 1-2',
'Blue 2-1', 'Blue 2-2', 'Green 2-1', 'Green 2-2', 'Red 2-1', 'Red 2-2',
'Case of flush'
]
fig, correlation_matrix = plot_pca_correlation_graph(x_norm,
features_name,
dimensions=(1, 2, 3, 4))
plt.show()
#K-means
#Elbow Method - SSE = Sum Squared Error
sse = []
for k in range(1, 15):
kmeans = KMeans(n_clusters=k, random_state=42)
kmeans.fit(x_norm)
sse.append(kmeans.inertia_)
kl = KneeLocator(range(1, 15), sse, curve='convex', direction='decreasing')
nbr_cluster = kl.elbow
from sklearn.preprocessing import StandardScaler
from mlxtend.plotting import plot_pca_correlation_graph
#%% load and standardize data
iris = datasets.load_iris()
X = iris.data
X_std = StandardScaler().fit_transform(X)
#%% specify feature names
feature_names = ['sepal length', 'sepal width', 'petal length', 'petal width']
#%%
# calculate the correlation matrix and
# create a correlation graph
fig, cor_mat = plot_pca_correlation_graph(X_std, \
feature_names, dimensions=(1, 2), \
figure_axis_size=10)
#%%
# show the numbers of the correlation
# matrix for the 4 features
print(cor_mat)
# console output:
# Dim 1 Dim 2
# sepal length -0.890169 -0.360830
# sepal width 0.460143 -0.882716
# petal length -0.991555 -0.023415
# petal width -0.964979 -0.064000
# %%
def test_pass_pca_corr():
X, y = iris_data()
plot_pca_correlation_graph(X, variables_names=['1', '2', '3', '4'])
ax.scatter(Ch[0], Ch[1], s=120)
ax.scatter(Dh[0], Dh[1], s=160)
ax.scatter(Eh[0], Eh[1], s=200)
ax.scatter(Fh[0], Fh[1], s=240)
ax.scatter(Gh[0], Gh[1], s=270)
ax.legend([('class ' + str(k + 1)) for k in range(7)])
plt.show()
plt.title(
'Projection by PCA of the zoo on the first 2 principal components')
plt.xlabel('PC1')
plt.ylabel('PC2')
## PCA circle (we just use a librairy to do this part)
figure, correlation_matrix = plot_pca_correlation_graph(
Y_norm, feature_names, dimensions=(1, 2), figure_axis_size=10)
plt.show()
## MDS
if (choice == 2 or choice == 3 or choice == 0):
Yh_norm = Y_norm.T
if (choice == 3 or choice == 0):
metric = []
for k in range(DIMENTION):
metric.append(
(abs(np.corrcoef(Yh_norm[k], Type)[0][1])
)) ## find correlation betweeen type and initial variables
Y_norm = Y_norm * metric #change the metric (that mean that's we encourage interspace between types.
S = np.dot(Y_norm, Y_norm.T)
def test_pass_pca_corr():
plot_pca_correlation_graph(X, ['1', '2', '3', '4'])