Neural Conformal Control for Time Series Forecasting

TL;DR

This paper presents a neural conformal prediction method for time series that improves adaptive coverage and interval consistency in non-stationary environments, utilizing multi-view data and monotonicity constraints.

Contribution

It introduces a neural network-based conformal prediction approach that enhances adaptivity, calibration, and consistency of prediction intervals in time series forecasting.

Findings

Significant improvements in coverage and probabilistic accuracy on real-world datasets.

The method uniquely combines good calibration with interval consistency.

Effective in non-stationary environments with auxiliary multi-view data.

Abstract

We introduce a neural network conformal prediction method for time series that enhances adaptivity in non-stationary environments. Our approach acts as a neural controller designed to achieve desired target coverage, leveraging auxiliary multi-view data with neural network encoders in an end-to-end manner to further enhance adaptivity. Additionally, our model is designed to enhance the consistency of prediction intervals in different quantiles by integrating monotonicity constraints and leverages data from related tasks to boost few-shot learning performance. Using real-world datasets from epidemics, electric demand, weather, and others, we empirically demonstrate significant improvements in coverage and probabilistic accuracy, and find that our method is the only one that combines good calibration with consistency in prediction intervals.