leidenalg
hdbscan
leidenalg | hdbscan | |
---|---|---|
1 | 6 | |
540 | 2,685 | |
- | 1.1% | |
8.0 | 5.8 | |
about 2 months ago | 11 days ago | |
Python | Jupyter Notebook | |
GNU General Public License v3.0 only | BSD 3-clause "New" or "Revised" License |
Stars - the number of stars that a project has on GitHub. Growth - month over month growth in stars.
Activity is a relative number indicating how actively a project is being developed. Recent commits have higher weight than older ones.
For example, an activity of 9.0 indicates that a project is amongst the top 10% of the most actively developed projects that we are tracking.
leidenalg
-
[OC] An interactive map of reddit built from 330 million user comments. 2023 update
- iGraph with Leidenalg uses C++ and exposes an interface to python
hdbscan
-
Introducing the Semantic Graph
A number of excellent topic modeling libraries exist in Python today. BERTopic and Top2Vec are two of the most popular. Both use sentence-transformers to encode data into vectors, UMAP for dimensionality reduction and HDBSCAN to cluster nodes.
- Hierarchical clustering algorithm
-
Introduction to K-Means Clustering
Working in spatial data science, I rarely find applications where k-means is the best tool. The problem is that it is difficult to know how many clusters you can expect on maps. Is it 5, 500, or 10,000? Here HDBSCAN [1] shines because it will cluster _and_ select the most suitable number of clusters, to cut the single linkage cluster tree.
[1]: https://github.com/scikit-learn-contrib/hdbscan
-
New clustering algorithms like DBSCAN and OPTICS?
You might be interested in HDBSCAN which has several implementations, but the python implelementation is commonly used. That implementation makes use of algorithmic changes to significantly improve the computational complexity. Some more recent variations on that include the gamma-linkage variant which is quite powerful.
-
DBSCAN ALternatives?
The OPTICS algorithm is in the latest versions of sklearn and is a reasonable alternative to DBSCAN -- it has much the same theoretical foundation, but can cope with variable density clusters better. If you are willing to step outside sklearn itself there is also HDBSCAN which is a hierarchical clustering version of DBSCAN and is in sklearn-contrib so should be compatible with an sklearn pipeline.
-
[D] Good algorithm for clustering big data (sentences represented as embeddings)?
Maybe use (H)DBScan which I think should work also for huge datasets. I don't think there is a ready to use clustering with unbuild cosine similarily metrics, and you also won't be able to precompute the 100k X 100k dense similarity matrix. The only way to go on this is to L2 normalize your embeddings, then the dot product will be the angular distance as a proxy to the cosine similarily. See also https://github.com/scikit-learn-contrib/hdbscan/issues/69
What are some alternatives?
kmodes - Python implementations of the k-modes and k-prototypes clustering algorithms, for clustering categorical data
faiss - A library for efficient similarity search and clustering of dense vectors.
communities - Library of community detection algorithms and visualization tools
Top2Vec - Top2Vec learns jointly embedded topic, document and word vectors.
HybridRenderingEngine - Clustered Forward/Deferred renderer with Physically Based Shading, Image Based Lighting and a whole lot of OpenGL.
Milvus - A cloud-native vector database, storage for next generation AI applications
fuzzy-c-means - A simple python implementation of Fuzzy C-means algorithm.
100DaysofMLCode - My journey to learn and grow in the domain of Machine Learning and Artificial Intelligence by performing the #100DaysofMLCode Challenge. Now supported by bright developers adding their learnings :+1:
Algorithms - Collection of algorithms in multiple programming languages.
RACplusplus - A high performance implementation of Reciprocal Agglomerative Clustering in C++
homemade-machine-learning - 🤖 Python examples of popular machine learning algorithms with interactive Jupyter demos and math being explained
word2vec - Automatically exported from code.google.com/p/word2vec