Preprocessing shapes time series forecasting model success
Are We Really Benchmarking Forecasting Models? The Impact of Preprocessing on Time Series Performance
Machine Learning
Summary
Forecasting methods often skip the important step of preparing data properly, which affects how well they work. The authors found that using different data preparation techniques can dramatically improve forecast accuracy, especially for simpler models that don't handle data changes internally. They tested many models and preparation methods on a large set of time series data and showed that tailoring preprocessing to each series can boost performance by up to nearly 90%. This means some simpler forecasting models could compete with the most advanced ones if their data is handled correctly.
What this means in practice
- •For data science teams: Improve forecasting accuracy by selecting the best data preprocessing method tailored to each time series before applying forecasting models.
- •For business intelligence analysts: Use optimized preprocessing to enable simpler forecasting models to deliver results comparable to complex models, saving computational resources in demand forecasting and planning.
Authors
Guilherme Afonso Galindo Padilha, Paulo Salgado Gomes de Mattos Neto, Rafael Menelau Oliveira e Cruz
Abstract
While established literature underscores the pivotal role of preprocessing in forecasting accuracy, this stage remains largely overlooked in current research. Modern benchmarks typically resort to simple scaling, failing to account for critical transformations required to address nonstationarity, such as differencing. This omission creates a significant structural preprocessing bias that favors models with built-in data treatments while obscuring the true potential of simpler architectures. We study this effect through a preprocessing-aware benchmark that evaluates 11 forecasting models across 16 reversible preprocessing pipelines on 29,000 M4 time series. Our results identify preprocessing as a key driver of forecasting performance. Optimizing preprocessing per series yields gains of approximately 27\% to 87\% across all evaluated models, with architectures lacking internalized preprocessing experiencing the most substantial improvements. This allows simpler architectures to become highly competitive with complex, state-of-the-art models in modern forecasting benchmarks. All resources and experimental results from this benchmark are stored in a comprehensive metadataset to support future metalearning tasks.