# \[ANN\] HorseML.jl v0.4.0
**URL:** https://discourse.julialang.org/t/ann-horseml-jl-v0-4-0/77181
**Category:** Package Announcements
**Created:** [February 28, 2022, 12:56pm UTC](https://discourse.julialang.org/t/ann-horseml-jl-v0-4-0/77181 "2022-02-28T12:56:53Z")
**Posts on this page:** 4
**Page:** 1
### Author:  [@QGMW22](https://discourse.julialang.org/u/QGMW22)
#### Post date: [February 28, 2022, 12:56pm UTC](https://discourse.julialang.org/t/ann-horseml-jl-v0-4-0/77181/1 "2022-02-28T12:56:53Z")
Announcing [HorseML](https://github.com/MommaWatasu/HorseML.jl) v0.4.0.
Clustering Models and data builder has been added.
This is a Clustering result:
Please see the [docs](https://mommawatasu.github.io/HorseML.jl/stable) for details.
---
### Author:  [@j\_u](https://discourse.julialang.org/u/j_u)
#### Post date: [February 28, 2022, 9:35pm UTC](https://discourse.julialang.org/t/ann-horseml-jl-v0-4-0/77181/2 "2022-02-28T21:35:48Z")
Hi and thanks. Would you consider adding / are you planning to add clustering technique that supports automatic determination of the number of clusters? I have only done basic reading on this topic so maybe HorseML.jl already provides such ability, if so would you please point me to the relevant part of the documentation? My experience with clustering is mostly based on K-means, K-medoids and DBSCAN from Custering.jl. For determining number of clusters I have been using the Elbow method, however, I always had to do some basic personal review and fine tune the results manually thus I am wondering if automatic determination of number of clusters is possible. [Not sure if this is relevant for this question but my clustering dataset is rather small (usually 32 points).]
---
### Author:  [@QGMW22](https://discourse.julialang.org/u/QGMW22)
#### Post date: [March 1, 2022, 1:47am UTC](https://discourse.julialang.org/t/ann-horseml-jl-v0-4-0/77181/3 "2022-03-01T01:47:31Z")
> [@j\_u](#):
>
> technique that supports automatic determination of the number of clusters?
Do you mean X-means?
Although I haven’t planned to add such technique, I’ll add it thinking it is necessary function.
---
### Author:  [@j\_u](https://discourse.julialang.org/u/j_u)
#### Post date: [March 1, 2022, 10:01pm UTC](https://discourse.julialang.org/t/ann-horseml-jl-v0-4-0/77181/4 "2022-03-01T22:01:08Z")
X-means is new to me. It sounds really good. Also, based on my experience with DBSCAN, my assumption is that HDBSCAN might be resulting with pretty good output. I am curious about Gaussian Mixture Model that is implemented in HorseML, I’ll try it ASAP, might not be immediately but … can’t wait to do it. Thanks again for the package.