Home
Creators
Search Recent Missing Random
Posts
Search DMs Popular Hash Lookup Tags Random
Importer
Import FAQ
Register Favorites Keys Review DMs Logout
Community
Telegram ThePornDude
Home Creators Posts Import Register Favorites Logout

Advanced Setups 21 - Jeroen Claus: HDBSCAN (Patreon)

Published:
2022-11-04 16:08:15
Imported:
2026-06
Tags:
haven't archived this post yet. have a subscription? use the importer!

Videos

Downloads

  • Entagma_Clustering_03_HDBSCAN.zip

Missing 1 full-res photo, 1 file, as this post has not yet been imported.

Content

Jeroen Claus covers another important clustering algorithm in the thrid part of his series on popular algorithms used in data analysis and visualisation. In this episode he’ll cover an improvement over DBSCAN, called HDBSCAN. Of course Jeroen goes over implementing it in Houdini using VEX.

Files

Previews only

Advanced Setups 21 - Jeroen Claus: HDBSCAN

Comments

Van

Fantastic buildup towards advanced understanding.

Matti Brandmaier

Great tutorial and a very interesting technique. Thank you for this! I hope advanced setups will continue with stuff regarding data analysis. I would like to share some findings I stumbled upon when trying to implement this technique working on larger datasets ( several 100k - millions of points ): using "int npts[] = nearpoints( 0, "@tree!=" + itoa( i@tree ), v@P, 10, 5 );" seems to greatly decrease in performance exponentially with more points to process. I guess this is due to the group call to exclude points with the same tree value. I got the (visually) same result by looking up couple of hundred points within a large radius and comparing the attributes in a foreach loop, this seemed to improve the performance by magnitudes. This is what my wrangle looks like: https://github.com/MattiBRNDMR/houdini_snippets/blob/main/pcfind_exclude_by_same_attribute Another bottleneck I've found when working with large data sets was the foeach point loop to promote the min distance etc. Rebuilding the foreach point loop with vex further accelerated the process: https://github.com/MattiBRNDMR/houdini_snippets/blob/main/get_min_on_a_dataset