Orbit: Probabilistic Forecast with Exponential Smoothing

TL;DR

This paper presents a refined Bayesian exponential smoothing model for time series forecasting, leveraging probabilistic programming to improve accuracy and flexibility, and compares it against other models on diverse datasets.

Contribution

It introduces a novel Bayesian exponential smoothing approach with enhancements like global trend and noise modeling, implemented using probabilistic programming languages.

Findings

The refined model outperforms traditional methods on benchmark datasets.

Probabilistic programming enables flexible and accurate time series modeling.

The approach effectively captures complex trend and noise structures.

Abstract

Time series forecasting is an active research topic in academia as well as industry. Although we see an increasing amount of adoptions of machine learning methods in solving some of those forecasting challenges, statistical methods remain powerful while dealing with low granularity data. This paper introduces a refined Bayesian exponential smoothing model with the help of probabilistic programming languages including Stan. Our model refinements include additional global trend, transformation for multiplicative form, noise distribution and choice of priors. A benchmark study is conducted on a rich set of time-series data sets for our models along with other well-known time series models.