ivis 1.2.2

Artificial neural network-driven visualization of high-dimensional data using triplets.

ivis

Implementation of the ivis algorithm as described in the paper Structure-preserving visualisation of high dimensional single-cell datasets.

This algorithm uses a siamese neural network trained on triplets to reduce the dimensionality of data to two dimensions for visualization. Each triplet is sampled from one of the k nearest neighbours as approximated by the Annoy library, with neighbouring points being pulled together and non-neighours being pushed away.

Unsupervised and supervised dimensionality reduction is supported.

Installation

Install the latest ivis releast from PyPi:

pip install ivis

Alternatively, you can install the development version from github:

git clone https://github.com/beringresearch/ivis
cd ivis
pip install -r requirements.txt --editable .

The following optional dependencies are needed if using the visualization callbacks while training the Ivis model:

• matplotlib
• seaborn

Upgrading

Ivis Python package is updated frequently! To upgrade, run:

pip install ivis --upgrade

Features

• Scalable: ivis is fast and easily extends to millions of observations and thousands of features.
• Versatile: numpy arrays, sparse matrices, and hdf5 files are supported out of the box, making it easy to apply ivis to heterogeneous problems including clustering and anomaly detection.
• Accurate: ivis excels at preserving both local and global features of a dataset. Often, ivis performs better at preserving global structure of the data than t-SNE, making it easy to visualise and interpret high-dimensional datasets.
• Generalisable: ivis supports addition of new data points to original embeddings via a transform method, making it easy to incorporate ivis into standard sklearn Pipelines.

And many more! See ivis readme for latest additions and examples.

Examples

from ivis import Ivis
from sklearn import datasets

iris = datasets.load_iris()
X = iris.data

model = Ivis(embedding_dims=2, k=15)

embeddings = model.fit_transform(X)

Ivis Universe

Ivis can be used in a wide variety of real-world applications. The Ivis Universe consists of packages that extend the core Ivis functionality.

• ivis-animate - visualise the Ivis learning process.
• ivis-explain - explain which features contribute the most to ivis embeddings.

Copyright 2019 Bering Limited