Replacing Classical Forecasting With Deep Learning Transformers

Author(s): Rashmi

Originally published on Towards AI.

Understanding the shift from classical ways to Transformer-based time series forecasting

Time-series forecasting has always been a critical component of finance, e-commerce, mobility, healthcare, manufacturing, and climate modeling. For decades, classical statistical models like ARIMA, SARIMA, ETS, and VAR dominated forecasting.

This article explores the transition from classical forecasting methods like ARIMA and ETS to deep learning transformers in time series forecasting. It discusses the limitations of classical models, the advantages of transformer architectures in capturing complex patterns and dependencies, use cases across different industries, and operational considerations. The piece highlights that while transformers excel in technical capabilities, classical methods remain effective for operational simplicity, particularly in instances of limited data availability.

Read the full blog for free on Medium.

Join thousands of data leaders on the AI newsletter. Join over 80,000 subscribers and keep up to date with the latest developments in AI. From research to projects and ideas. If you are building an AI startup, an AI-related product, or a service, we invite you to consider becoming a sponsor.

Published via Towards AI