Merlion
A Machine Learning Library for Time Series
Meet Merlion
Merlion is an open-source Python library for time series intelligence. Merlion provides an end-to-end machine learning framework that includes loading and transforming data, building and training models, post-processing model outputs, and evaluating model performance. It supports various time series learning tasks, including forecasting, anomaly detection, and change-point detection for both univariate and multivariate time series. This library solves a range of problems by providing engineers and researchers a one-stop solution to rapidly develop models for their specific time series needs, and benchmark them across multiple time-series datasets. Instead of having to learn and deploy multiple tools, you can do it all within a single, powerful framework.
Data Layer
Standardized loaders for popular open source time series datasets. Support for many standard data pre-processing transforms.
Models
A library of diverse models, unified under a shared interface. We provide "default" models for new users, and fully configurable individual models for advanced users.
AutoML
Easy-to-use autoML for hyperparameter tuning and model selection.
Ensembles
Transparent support for creating heterogeneous ensembles of anomaly detection or forecasting models.
Post-Processing
Practical, industry-inspired post-processing rules for anomaly detectors that make anomaly scores more interpretable, while also reducing the number of false positives.
Evaluation and Benchmarking
Flexible evaluation pipelines to simulate the live deployment of a model in production. Scripts to benchmark model performance on multiple datasets. Get started on GitHub.
Latest Publications
Read through our latest paper and blog to discover Merlion's functionalities and how it compares to other time-series libraries.
Meet Merlion: An End-to-End Easy-to-Use Machine Learning Library for Time Series Applications
read on blog@article{bhatnagar2021merlion,
title={Merlion: A Machine Learning Library for Time Series},
author={Aadyot Bhatnagar and Paul Kassianik and Chenghao Liu and Tian Lan and Wenzhuo Yang
and Rowan Cassius and Doyen Sahoo and Devansh Arpit and Sri Subramanian and Gerald Woo
and Amrita Saha and Arun Kumar Jagota and Gokulakrishnan Gopalakrishnan and Manpreet Singh
and K C Krithika and Sukumar Maddineni and Daeki Cho and Bo Zong and Yingbo Zhou
and Caiming Xiong and Silvio Savarese and Steven Hoi and Huan Wang},
year={2021},
eprint={2109.09265},
archivePrefix={arXiv},
primaryClass={cs.LG}
}