def _rank_documents(self, doc_list, threshold, tolerance):
n = len(doc_list)
#Initialises the adjacency matrix
adjacency_matrix = numpy.zeros([n, n])
degree = numpy.zeros([n])
scores = numpy.zeros([n])
for i, documenti in enumerate(doc_list):
for j, documentj in enumerate(doc_list):
adjacency_matrix[i][j] = cosine(documenti.fv,
documentj.fv,
distance=False)
if adjacency_matrix[i][j] > threshold:
adjacency_matrix[i][j] = 1.0
degree[i] += 1
else:
adjacency_matrix[i][j] = 0
for i in xrange(n):
for j in xrange(n):
if degree[i] == 0: degree[i] = 1.0 #at least similat to itself
adjacency_matrix[i][j] = adjacency_matrix[i][j] / degree[i]
scores = self.power_method(adjacency_matrix, tolerance)
for i in xrange(0, n):
doc_list[i].dist = scores[i]
return doc_list
def _rank_documents(self, doc_list, threshold, tolerance):
n = len(doc_list)
# Initialises the adjacency matrix
adjacency_matrix = numpy.zeros([n, n])
degree = numpy.zeros([n])
scores = numpy.zeros([n])
for i, documenti in enumerate(doc_list):
for j, documentj in enumerate(doc_list):
adjacency_matrix[i][j] = cosine(documenti.fv, documentj.fv, distance=False)
if adjacency_matrix[i][j] > threshold:
adjacency_matrix[i][j] = 1.0
degree[i] += 1
else:
adjacency_matrix[i][j] = 0
for i in xrange(n):
for j in xrange(n):
if degree[i] == 0:
degree[i] = 1.0 # at least similat to itself
adjacency_matrix[i][j] = adjacency_matrix[i][j] / degree[i]
scores = self.power_method(adjacency_matrix, tolerance)
for i in xrange(0, n):
doc_list[i].dist = scores[i]
return doc_list
def _calculate_document_scores(self):
'''
Calculates the cosine similarity between a document and the centroid.
It attaches the calculated distance on the document.
'''
for id, document in self.documents.iteritems():
dist = cosine(self.centroid, document.fv)
self.documents[id].dist = dist
def _calculate_document_scores(self):
"""
Calculates the cosine similarity between a document and the centroid.
It attaches the calculated distance on the document.
"""
for id, document in self.documents.iteritems():
dist = cosine(self.centroid, document.fv)
self.documents[id].dist = dist
