def __get_variance_increase_distance(self, entry):
"""!
@brief Calculates variance increase distance between current and specified clusters.
@param[in] entry (cfentry): Clustering feature to which distance should be obtained.
@return (double) Variance increase distance.
"""
linear_part_12 = list_math_addition(self.linear_sum, entry.linear_sum)
variance_part_first = (self.square_sum + entry.square_sum) - \
2.0 * sum(list_math_multiplication(linear_part_12, linear_part_12)) / (self.number_points + entry.number_points) + \
(self.number_points + entry.number_points) * sum(list_math_multiplication(linear_part_12, linear_part_12)) / (self.number_points + entry.number_points)**2.0
linear_part_11 = sum(
list_math_multiplication(self.linear_sum, self.linear_sum))
variance_part_second = -(self.square_sum -
(2.0 * linear_part_11 / self.number_points) +
(linear_part_11 / self.number_points))
linear_part_22 = sum(
list_math_multiplication(entry.linear_sum, entry.linear_sum))
variance_part_third = -(
entry.square_sum - (2.0 / entry.number_points) * linear_part_22 +
entry.number_points *
(1.0 / entry.number_points**2.0) * linear_part_22)
return (variance_part_first + variance_part_second +
variance_part_third)
def get_radius(self):
"""!
@brief Calculates radius of cluster that is represented by the entry.
@details It's calculated once when it's requested after the last changes.
@return (double) Radius of cluster that is represented by the entry.
"""
if (self.__radius is not None):
return self.__radius;
centroid = self.get_centroid();
radius_part_1 = self.square_sum;
radius_part_2 = 0.0;
radius_part_3 = 0.0;
if (type(centroid) == list):
radius_part_2 = 2.0 * sum(list_math_multiplication(self.linear_sum, centroid));
radius_part_3 = self.number_points * sum(list_math_multiplication(centroid, centroid));
else:
radius_part_2 = 2.0 * self.linear_sum * centroid;
radius_part_3 = self.number_points * centroid * centroid;
self.__radius = ( (1.0 / self.number_points) * (radius_part_1 - radius_part_2 + radius_part_3) ) ** 0.5;
return self.__radius;
def get_radius(self):
"""!
@brief Calculates radius of cluster that is represented by the entry.
@details It's calculated once when it's requested after the last changes.
@return (double) Radius of cluster that is represented by the entry.
"""
if (self.__radius is not None):
return self.__radius
centroid = self.get_centroid()
radius_part_1 = self.square_sum
radius_part_2 = 0.0
radius_part_3 = 0.0
if (type(centroid) == list):
radius_part_2 = 2.0 * sum(
list_math_multiplication(self.linear_sum, centroid))
radius_part_3 = self.number_points * sum(
list_math_multiplication(centroid, centroid))
else:
radius_part_2 = 2.0 * self.linear_sum * centroid
radius_part_3 = self.number_points * centroid * centroid
self.__radius = ((1.0 / self.number_points) *
(radius_part_1 - radius_part_2 + radius_part_3))**0.5
return self.__radius
def __minimum_noiseless_description_length(self, clusters, centers):
"""!
@brief Calculates splitting criterion for input clusters using minimum noiseless description length criterion.
@param[in] clusters (list): Clusters for which splitting criterion should be calculated.
@param[in] centers (list): Centers of the clusters.
@return (double) Returns splitting criterion in line with bayesian information criterion.
Low value of splitting cretion means that current structure is much better.
@see __bayesian_information_criterion(clusters, centers)
"""
scores = [0.0] * len(clusters);
W = 0.0;
K = len(clusters);
N = 0.0;
sigma_sqrt = 0.0;
alpha = 0.9;
betta = 0.9;
for index_cluster in range(0, len(clusters), 1):
for index_object in clusters[index_cluster]:
delta_vector = list_math_subtraction(self.__pointer_data[index_object], centers[index_cluster]);
delta_sqrt = sum(list_math_multiplication(delta_vector, delta_vector));
W += delta_sqrt;
sigma_sqrt += delta_sqrt;
N += len(clusters[index_cluster]);
if (N - K != 0):
W /= N;
sigma_sqrt /= (N - K);
sigma = sigma_sqrt ** 0.5;
for index_cluster in range(0, len(clusters), 1):
Kw = (1.0 - K / N) * sigma_sqrt;
Ks = ( 2.0 * alpha * sigma / (N ** 0.5) ) + ( (alpha ** 2.0) * sigma_sqrt / N + W - Kw / 2.0 ) ** 0.5;
U = W - Kw + 2.0 * (alpha ** 2.0) * sigma_sqrt / N + Ks;
Z = K * sigma_sqrt / N + U + betta * ( (2.0 * K) ** 0.5 ) * sigma_sqrt / N;
if (Z == 0.0):
scores[index_cluster] = float("inf");
else:
scores[index_cluster] = Z;
else:
scores = [float("inf")] * len(clusters);
return sum(scores);
def __get_average_inter_cluster_distance(self, entry):
"""!
@brief Calculates average inter cluster distance between current and specified clusters.
@param[in] entry (cfentry): Clustering feature to which distance should be obtained.
@return (double) Average inter cluster distance.
"""
linear_part_distance = sum(list_math_multiplication(self.linear_sum, entry.linear_sum));
return ( (entry.number_points * self.square_sum - 2.0 * linear_part_distance + self.number_points * entry.square_sum) / (self.number_points * entry.number_points) ) ** 0.5;
def __get_variance_increase_distance(self, entry):
"""!
@brief Calculates variance increase distance between current and specified clusters.
@param[in] entry (cfentry): Clustering feature to which distance should be obtained.
@return (double) Variance increase distance.
"""
linear_part_12 = list_math_addition(self.linear_sum, entry.linear_sum);
variance_part_first = (self.square_sum + entry.square_sum) - \
2.0 * sum(list_math_multiplication(linear_part_12, linear_part_12)) / (self.number_points + entry.number_points) + \
(self.number_points + entry.number_points) * sum(list_math_multiplication(linear_part_12, linear_part_12)) / (self.number_points + entry.number_points)**2.0;
linear_part_11 = sum(list_math_multiplication(self.linear_sum, self.linear_sum));
variance_part_second = -( self.square_sum - (2.0 * linear_part_11 / self.number_points) + (linear_part_11 / self.number_points) );
linear_part_22 = sum(list_math_multiplication(entry.linear_sum, entry.linear_sum));
variance_part_third = -( entry.square_sum - (2.0 / entry.number_points) * linear_part_22 + entry.number_points * (1.0 / entry.number_points ** 2.0) * linear_part_22 );
return (variance_part_first + variance_part_second + variance_part_third);
def __get_average_inter_cluster_distance(self, entry):
"""!
@brief Calculates average inter cluster distance between current and specified clusters.
@param[in] entry (cfentry): Clustering feature to which distance should be obtained.
@return (double) Average inter cluster distance.
"""
linear_part_distance = sum(
list_math_multiplication(self.linear_sum, entry.linear_sum))
return ((entry.number_points * self.square_sum -
2.0 * linear_part_distance +
self.number_points * entry.square_sum) /
(self.number_points * entry.number_points))**0.5
def __get_average_intra_cluster_distance(self, entry):
"""!
@brief Calculates average intra cluster distance between current and specified clusters.
@param[in] entry (cfentry): Clustering feature to which distance should be obtained.
@return (double) Average intra cluster distance.
"""
linear_part_first = list_math_addition(self.linear_sum, entry.linear_sum);
linear_part_second = linear_part_first;
linear_part_distance = sum(list_math_multiplication(linear_part_first, linear_part_second));
general_part_distance = 2.0 * (self.number_points + entry.number_points) * (self.square_sum + entry.square_sum) - 2.0 * linear_part_distance;
return (general_part_distance / ( (self.number_points + entry.number_points) * (self.number_points + entry.number_points - 1.0) )) ** 0.5;
def get_diameter(self):
"""!
@brief Calculates diameter of cluster that is represented by the entry.
@details It's calculated once when it's requested after the last changes.
@return (double) Diameter of cluster that is represented by the entry.
"""
if (self.__diameter is not None):
return self.__diameter;
diameter_part = 0.0;
if (type(self.linear_sum) == list):
diameter_part = self.square_sum * self.number_points - 2.0 * sum(list_math_multiplication(self.linear_sum, self.linear_sum)) + self.square_sum * self.number_points;
else:
diameter_part = self.square_sum * self.number_points - 2.0 * self.linear_sum * self.linear_sum + self.square_sum * self.number_points;
self.__diameter = ( diameter_part / (self.number_points * (self.number_points - 1)) ) ** 0.5;
return self.__diameter;
def get_diameter(self):
"""!
@brief Calculates diameter of cluster that is represented by the entry.
@details It's calculated once when it's requested after the last changes.
@return (double) Diameter of cluster that is represented by the entry.
"""
if (self.__diameter is not None):
return self.__diameter
diameter_part = 0.0
if (type(self.linear_sum) == list):
diameter_part = self.square_sum * self.number_points - 2.0 * sum(
list_math_multiplication(self.linear_sum, self.linear_sum)
) + self.square_sum * self.number_points
else:
diameter_part = self.square_sum * self.number_points - 2.0 * self.linear_sum * self.linear_sum + self.square_sum * self.number_points
self.__diameter = (diameter_part / (self.number_points *
(self.number_points - 1)))**0.5
return self.__diameter
def __get_average_intra_cluster_distance(self, entry):
"""!
@brief Calculates average intra cluster distance between current and specified clusters.
@param[in] entry (cfentry): Clustering feature to which distance should be obtained.
@return (double) Average intra cluster distance.
"""
linear_part_first = list_math_addition(self.linear_sum,
entry.linear_sum)
linear_part_second = linear_part_first
linear_part_distance = sum(
list_math_multiplication(linear_part_first, linear_part_second))
general_part_distance = 2.0 * (
self.number_points + entry.number_points
) * (self.square_sum + entry.square_sum) - 2.0 * linear_part_distance
return (general_part_distance /
((self.number_points + entry.number_points) *
(self.number_points + entry.number_points - 1.0)))**0.5