MATLAB wrapper for umap

This repo provides a simple wrapper for umap

Installation

0. Install a required dependency using git clone git@github.com:sg-s/srinivas.gs_mtools.git
1. Add that folder to your MATLAB path. Don't add subfolders.
2. Install umap using conda install -c conda-forge umap-learn (paste this in your shell, not your MATLAB prompt)
3. Install h5py using conda install h5py
4. (optional) Install condalab
5. (optional) Run conda.init to configure condalab in MATLAB

Use

u = umap();
R = u.fit(X);

Parameters and options

Many of the parameters and options in umap are exposed in the object, and you can change these directly from MATLAB. For example:

u = umap

  umap with properties:

             n_neighbors: 15
            n_components: 2
                  metric: 'euclidean'
           learning_rate: 1
                min_dist: 0.1000
                  spread: 1
        set_op_mix_ratio: 1
      local_connectivity: 1
      repulsion_strength: 1
    negative_sample_rate: 5
    transform_queue_size: 4
      target_n_neighbors: -1
           target_weight: 0.5000
          transform_seed: 42


u.n_neighbors = 10;
u.metric = 'precomputed';

License

GPL