EDA package 1.1.30
This release also includes Datagrok dimensionality reduction toolkit, which supports combination of multiple columns of different types (molecules, Sequences, numbers…) with corresponding distance functions and light-speed performance with use of WebGPU.
You can read more about the dimensionality reduction capabilities of Datagrok here.
To run dimensionality reduction:
- Open dataframe
- Go to Top Menu > ML > Dimensionality Reduction…
- Specify
Columns - For each column, specify
distance functionandweight - Choose desired method (UMAP or t-SNE) along with their hyper-parameters
- Choose desired post-processing (clustering of embeddings)
