Andrei Novikov

Results 44 issues of Andrei Novikov

**Description** Coverity Scan should be introduced for pyclustering library. In line policy, it shouldn't be run every commit therefore there are several variants how to do that:: 1. Nightly cron...

Continuous Integration

**Introduction** Distinguish UPGMA and WPGMA for Python and C++ parts of the library. **Description** Currently, UPGMA is implemented and called as average linkage. - [ ] Documentation revise; - [...

Enhancement
Documentaion
Good First Issue

**Introduction** 'random_state' parameter should be introduced in order to provide control to random seed value that is used during the processing. **Description** - [ ] `random_state` parameter should be passed...

Enhancement
Good First Issue

**Introduction** As a tester or developer, I want to have correct answer for Hepta FCPS sample so that I can use it in test scenarios to improve quality of the...

Testing
Good First Issue

**Introduction** Current python implementation `pyclustering.cluster.clique` is not optimal in case of high dimensional data. See complaint: #634 **Description** Algorithm should work for high-dimensional data. No need to build blocks physically,...

Optimization

**Introduction** The current problem is following: in case of incorrect input data or due to some other reason (even unexpected) - C++ code throws exception that is not captured by...

Enhancement
Good First Issue

**Introduction** STING (a STatistical INformation Grid approach) clustering algorithm. The general idea is to divide spatial aria into rectangular cells at different levels of resolution which forms tree structure. Statistical...

Enhancement
Good First Issue

**Introduction** Almost all objects are returned from C++ pyclustering to python pyclustering using `pyclustering_package`. See files `ccore/include/interface/pyclustering_package.hpp` and `ccore/src/interface/pyclustering_package.cpp`. **Description** In order to return error message new type of data...

Enhancement
Good First Issue

**Introduction** DBSCAN algorithm should be able to accept distance metric in the same way as K-Means, K-Medians, etc. **Description** - [ ] Introduce new optional argument `metric` (`distance_metric` type) under...

Enhancement
Testing
Documentaion

**Introduction** Support custom distance metric for KD-tree in order to provide way to use Euclidean, Square Euclidean, Manhattan, Chebyshev and other metrics. **Description** - [ ] Introduce optional parameter `metric`...

Enhancement
Testing
Documentaion