annoviko /
pyclustering
| @@ 90-120 (lines=31) @@ | ||
| 87 | self.__ccore = ccore_library.workable() |
|
| 88 | ||
| 89 | ||
| 90 | def process(self): |
|
| 91 | """! |
|
| 92 | @brief Performs cluster analysis in line with rules of K-Medians algorithm. |
|
| 93 | ||
| 94 | @remark Results of clustering can be obtained using corresponding get methods. |
|
| 95 | ||
| 96 | @see get_clusters() |
|
| 97 | @see get_medians() |
|
| 98 | ||
| 99 | """ |
|
| 100 | ||
| 101 | if self.__ccore is True: |
|
| 102 | ccore_metric = metric_wrapper.create_instance(self.__metric) |
|
| 103 | ||
| 104 | self.__clusters = wrapper.kmedians(self.__pointer_data, self.__medians, self.__tolerance, ccore_metric.get_pointer()) |
|
| 105 | self.__medians = self.__update_medians() |
|
| 106 | ||
| 107 | else: |
|
| 108 | changes = float('inf')
|
|
| 109 | ||
| 110 | # Check for dimension |
|
| 111 | if len(self.__pointer_data[0]) != len(self.__medians[0]): |
|
| 112 | raise NameError('Dimension of the input data and dimension of the initial medians must be equal.')
|
|
| 113 | ||
| 114 | while changes > self.__tolerance: |
|
| 115 | self.__clusters = self.__update_clusters() |
|
| 116 | updated_centers = self.__update_medians() |
|
| 117 | ||
| 118 | changes = max([self.__metric(self.__medians[index], updated_centers[index]) for index in range(len(updated_centers))]) |
|
| 119 | ||
| 120 | self.__medians = updated_centers |
|
| 121 | ||
| 122 | ||
| 123 | def get_clusters(self): |
|
| @@ 133-161 (lines=29) @@ | ||
| 130 | self.__ccore = ccore_library.workable() |
|
| 131 | ||
| 132 | ||
| 133 | def process(self): |
|
| 134 | """! |
|
| 135 | @brief Performs cluster analysis in line with rules of K-Medoids algorithm. |
|
| 136 | ||
| 137 | @remark Results of clustering can be obtained using corresponding get methods. |
|
| 138 | ||
| 139 | @see get_clusters() |
|
| 140 | @see get_medoids() |
|
| 141 | ||
| 142 | """ |
|
| 143 | ||
| 144 | if self.__ccore is True: |
|
| 145 | ccore_metric = metric_wrapper.create_instance(self.__metric) |
|
| 146 | ||
| 147 | self.__clusters = wrapper.kmedoids(self.__pointer_data, self.__medoid_indexes, self.__tolerance, ccore_metric.get_pointer(), self.__data_type) |
|
| 148 | self.__medoid_indexes = self.__update_medoids() |
|
| 149 | ||
| 150 | else: |
|
| 151 | changes = float('inf')
|
|
| 152 | ||
| 153 | stop_condition = self.__tolerance |
|
| 154 | ||
| 155 | while changes > stop_condition: |
|
| 156 | self.__clusters = self.__update_clusters() |
|
| 157 | update_medoid_indexes = self.__update_medoids() |
|
| 158 | ||
| 159 | changes = max([self.__distance_calculator(self.__medoid_indexes[index], update_medoid_indexes[index]) for index in range(len(update_medoid_indexes))]) |
|
| 160 | ||
| 161 | self.__medoid_indexes = update_medoid_indexes |
|
| 162 | ||
| 163 | ||
| 164 | def get_clusters(self): |
|
