search

eren-ck / st_dbscan

ST-DBSCAN: Simple and effective tool for spatial-temporal clustering
MIT License
128 stars 25 forks source link

density factor implementation #6

Closed priamai closed 3 years ago

priamai commented 3 years ago

Hi there, I was reading the original paper and your implementation and correct if I am wrong but there is no provision for the density factor thus there will be issues identifying adjacent clusters. Unless this is implemented in the standard DBSCAN algorithm in SKLEARN but I can't find any info in there either.

eren-ck commented 3 years ago

Hello, yes, the density factor as described in the paper [1] is currently not implemented. However, we assume that if you use the fit_frame_split as described in the paper [2], that to some degree different densities over time will also be supported. However, not within the same partitioned frame. A potential solution for this is to simply replace the algorithm in a fork with the HDBSCAN algorithm (link) which is able to capture different densities.

I hope this helps you.

Cheers, Eren

  1. Birant, Derya, and Alp Kut. "ST-DBSCAN: An algorithm for clustering spatial–temporal data." Data & Knowledge Engineering 60.1 (2007): 208-221.
  2. Peca, I., Fuchs, G., Vrotsou, K., Andrienko, N. V., & Andrienko, G. L. (2012). Scalable Cluster Analysis of Spatial Events. In EuroVA@ EuroVis.
priamai commented 3 years ago

Awesome thanks for the pointers, I am reading them.