pyclustering
pyclustering copied to clipboard
annoviko
pyclustering is a Python, C++ data mining library.
**Introduction** There is request to obtain implementation of Sparse Subspace Clustering algorithm. **Description** Link to article 'Sparse Subspace Clustering: Algorithm, Theory, and Applications': https://arxiv.org/pdf/1203.1005.pdf
**Introduction** In case of big SOM network sizes memory consumption is significantly big because of O(N*M), where N - amount of rows and M - amount of columns. **Description** -...
**Introduction** Currently, adjacent matrix or adjacent lists are created for any type of network structure. **Description** Memory and performance optimization should be done for 'all-to-all' network structure. Exception should be...
**Introduction** Currently, adjacent matrix or adjacent lists are created for any type of network structure. **Description** Memory and performance optimization should be done for 'all-to-all' network structure. Exception should be...
Is there a way to know the progress (like 50% complete, 20min ETA etc) for each of these functions? ``` cure_instance = cure(X, k, ccore=True); cure_instance.process(); clusters = cure_instance.get_clusters(); ```...
**Introduction** Clustering algorithm BANG is implemented using python. This algorithm should be supported by CCORE (C/C++ implementation) to increase performance of the algorithm. **Description** Following task should done in scope...
**Introduction** Regression tests are required for clustering algorithms. **Description** Requirement for regression testing - Each algorithm should be tested from 'pyclustering.cluster' module. - Regression time should be specified. - Various...
**Introduction** In case of user-defined metric, algorithms doesn't use core to avoid crash on some platforms. **Description** Investigation is required. See test: 'testUserDefinedMetric'. Test location: **pyclustering/utils/tests/integration/it_metric.py** ``` Ran 6 tests...
**Introduction** Sequential algorithm BSAS can be used for stream processing. Current implementation requires whole dataset at once and adding new point requires to repeat processing again. Therefore new method is...
**Introduction** Lawrence Shampine, Marilyn Gordon ODE solver implementation. **Description** C/C++ implementation with ccore.differential interface. Add usage of this algorithm in HHN, LEGION, Sync, and other networks that uses differential equations.