Skip to content

Added option for variance #1

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Open
matteosavarese95 wants to merge 1 commit into
base: master
Choose a base branch
from
Open

Added option for variance #1

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
59 changes: 47 additions & 12 deletions OpenMORe/clustering.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -101,6 +101,8 @@ def __init__(self, X, *dictionary):
self._k = 2
#Initialize the number of PCs to retain in each cluster:
self._nPCs = 2
#Initialize variance to preserve
self._var = 0.95
#Set the initialization method:
self._method = 'uniform' #Available options: 'KMEANS' or 'RANDOM'
#Set the (eventual) corrector for the rec error computation:
Expand All @@ -122,6 +124,9 @@ def __init__(self, X, *dictionary):

self._postKNN = False
self._neighborsNum = 0

# Stopping method: var or nPCs
self._stop = 'nPCs' # By default the stopping method is by selecting number of PCs

if dictionary:
settings = dictionary[0]
Expand All @@ -135,17 +140,42 @@ def __init__(self, X, *dictionary):
print("\tIt will be automatically set equal to: 2.")
print("\tYou can ignore this warning if the number of clusters (k) has been assigned later via setter.")
print("\tOtherwise, please check the conditions which must be satisfied by the input in the detailed documentation.")

try:
self._nPCs = settings["number_of_eigenvectors"]
if self._nPCs <= 0 or self._nPCs >= self.X.shape[1]:
self._stop = settings["stopping_method"]
if self._stop not in ['nPCs', 'var']:
raise Exception
except:
self._nPCs = int(self.X.shape[1]/2)
warnings.warn("An exception occured with regard to the input value for the number of PCs. It could be not acceptable, or not given to the dictionary.")
print("\tIt will be automatically set equal to: X.shape[1]-1.")
print("\tYou can ignore this warning if the number of PCs has been assigned later via setter.")
self._stop = 'nPCs'
warnings.warn("An exception occured with regard to the input value for the stopping method. It could be not acceptable, or not given to the dictionary.")
print("\tIt will be automatically set equal to: 'nPCs'.")
print("\tYou can ignore this warning if the stopping method has been assigned later via setter.")
print("\tOtherwise, please check the conditions which must be satisfied by the input in the detailed documentation.")

if self._stop == 'nPCs':
try:
self._nPCs = settings["number_of_eigenvectors"]
if self._nPCs <= 0 or self._nPCs >= self.X.shape[1]:
raise Exception
except:
self._nPCs = int(self.X.shape[1]/2)
warnings.warn("An exception occured with regard to the input value for the number of PCs. It could be not acceptable, or not given to the dictionary.")
print("\tIt will be automatically set equal to: X.shape[1]-1.")
print("\tYou can ignore this warning if the number of PCs has been assigned later via setter.")
print("\tOtherwise, please check the conditions which must be satisfied by the input in the detailed documentation.")

elif self._stop == 'var':
try:
self._var = settings["variance_threshold"]
if isinstance(self._var, float) and self._var > 1:
raise Exception
except:
self._var = 0.95
warnings.warn("An exception occured with regard to the input value for the variance threshold. It could be not acceptable, or not given to the dictionary.")
print("\tIt will be automatically set equal to: 0.95.")
print("\tYou can ignore this warning if the variance threshold has been assigned later via setter.")
print("\tOtherwise, please check the conditions which must be satisfied by the input in the detailed documentation.")

try:
self._center = settings["center"]
if not isinstance(self._center, bool):
Expand Down Expand Up @@ -729,6 +759,8 @@ def fit(self):
correction_ = np.zeros((rows, self._k), dtype=float)
scores_factor = np.zeros((rows, self._k), dtype=float)
# Iterate
# Get local eigenvectors and eigenvalues
local_eigenvectors = [None] * self._k # This is now saved as output
while(iteration < iter_max):
sq_rec_oss = np.zeros((rows, cols), dtype=float)
sq_rec_err = np.zeros((rows, self._k), dtype=float)
Expand All @@ -751,13 +783,16 @@ def fit(self):
#perform PCA in the cluster, centering and scaling can be avoided
#because the observations are already standardized
local_model = model_order_reduction.PCA(cluster)
local_model.to_center = False
local_model.to_center = True # Coherent with Parente's code
local_model.to_scale = False
if not self._adaptive:
if self._stop == "nPCs":
local_model.eigens = self._nPCs
else:
elif self._stop == "var":
local_model._threshold_var = self._var
local_model.set_PCs()
print("Number of PCs: {}".format(local_model.eigens))
modes = local_model.fit()
local_eigenvectors[ii] = np.copy(modes[0])
#create the centroids (medoids or medianoids, respectively) matrix
C_mat = np.matlib.repmat(centroids, rows, 1)
#compute the rec error for the considered cluster
Expand Down Expand Up @@ -867,7 +902,7 @@ def fit(self):
# Consider only statistical meaningful groups of points: if there are <2 points
#in a cluster, delete it because it's not statistically meaningful
idx = self.merge_clusters(self.X_tilde, idx)
return idx
return idx, local_eigenvectors


class fpca(lpca):
Expand Down Expand Up @@ -1653,4 +1688,4 @@ def fitApprox(self):

index = modelK.fit()

return index
return index