NeuralProphet: Better Time Series Forecasting (Currently in Beta)

Full Article: [2111.15397] NeuralProphet: Explainable Forecasting at Scale (arxiv.org)
Citation: @misc{triebe2021neuralprophet, title={NeuralProphet: Explainable Forecasting at Scale}, author={Oskar Triebe and Hansika Hewamalage and Polina Pilyugina and Nikolay Laptev and Christoph Bergmeir and Ram Rajagopal}, year={2021}, eprint={2111.15397}, archivePrefix={arXiv}, primaryClass={cs.LG}}

Time series forecasting plays a crucial role in decision-making processes across various industries. Whether it’s predicting stock prices, weather patterns, or sales trends, the ability to accurately forecast future events is invaluable. However, traditional methods often require deep domain knowledge and extensive expertise to achieve accurate results. Enter NeuralProphet, an innovative, human-centered framework designed to bridge the gap between classic time series models and modern machine learning techniques.

What is NeuralProphet?

NeuralProphet is a powerful, yet user-friendly, framework for interpretable time series forecasting. Built on top of PyTorch, NeuralProphet combines the strengths of neural networks with traditional time-series algorithms. It draws inspiration from Facebook’s Prophet model and AR-Net (Auto-Regression Network), providing a hybrid…