[pyclustering.cluster] STING algorithm

[annoviko]

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 information of each cell is calculated (mean, count, standard deviation, min, max) and type of distribution is also calculated (normal, uniform). After that regions queries are processed.

Resources

1. Lecture: https://www.coursera.org/learn/cluster-analysis/lecture/vuf0F/5-5-sting-a-statistical-information-grid-approach
2. Article: STING: A statistical information grid approach to spatial data mining.

Description General functionality that should be provided by the algorithm:

1. Parallel implementation of query processing.

2. Following interface should be used:

   class sting:
   def process():
    pass;
   
   def get_clusters():
     pass;

3. Visualization of the tree.
4. Visualization of the queries.

What should be also done:

• Unit-tests for the algorithm must be implemented: 'pyclustering/cluster/tests/unit/ut_sting.py'
• Examples for the algorithm must be implemented: 'pyclustering/cluster/examples/sting_examples.py'
• Doxygen documentation must be written for each public method, class, variable and it should be switched on in 'docs/doxygen_conf_pyclustering'

[tongliuTL]

Wonder why this algorithm was not developed.

[annoviko]

@tongliuTL , the answer is quite simple, because I am the only developer in that open-source project and I am working on it during free-time only.