clairvoyance2 0.0.0

clairvoyance2: a Unified Toolkit for Medical Time Series

clairvoyance2

clairvoyance2: a Unified Toolkit for Medical Time Series

⚠️ The library is in pre-alpha / dev. API will change without warning.

clairvoyance2 is a library that unifies time series tasks for the medicine and healthcare use case. It provides tools for manipulating multi-dimensional time series, as well as static data, and implements models for: time series prediction, individualized treatment effects estimation, time-to-event analysis (upcoming), and model interpretability (upcoming). clairvoyance2 is primarily focussed on machine learning models.

Installation

pip install git+https://github.com/vanderschaarlab/clairvoyance2.git

Models

Model	Status
Prediction (Forecasting)
RNN	✅ Implemented
Seq2Seq	✅ Implemented
NeuralLaplace	🔲 Planned
Imputation
{f,b}fill & Mean	✅ Implemented
HyperImpute	🔲
Individualized Treatment Effects
CRN	✅ Implemented
SyncTwin	⚙️ Experimental
TE-CDE	🔲 Planned
Time-to-event Analysis
Dynamic DeepHit	🔲 Planned
Interpretability
DynaMask	🔲 Planned

Tutorials

• 📔 Basic Usage
• 📔 Example: CRN
• 📔 Example: SyncTwin

Contact

If you wish to reach about to us specifically about clairvoyance2 (bugs, suggestions, problems, ...) please message Evgeny on LinkedIn for now, until we set up an official communication channel.