StaDRe and StaDRo: Reliability and Robustness Estimation of ML-based Forecasting using Statistical Distance Measures

TL;DR

This paper introduces StaDRe and StaDRo, novel statistical distance-based measures for estimating the reliability and robustness of ML forecasting models, especially under distributional shifts in time series data.

Contribution

It extends SafeML to time series forecasting by proposing SDD-based reliability and robustness measures, linking dataset shifts to model KPIs.

Findings

Effective detection of distributional shifts in time series forecasts

Correlation between dataset SDD and model KPIs established

Improved reliability and robustness estimation methods demonstrated

Abstract

Reliability estimation of Machine Learning (ML) models is becoming a crucial subject. This is particularly the case when such \mbox{models} are deployed in safety-critical applications, as the decisions based on model predictions can result in hazardous situations. In this regard, recent research has proposed methods to achieve safe, \mbox{dependable}, and reliable ML systems. One such method consists of detecting and analyzing distributional shift, and then measuring how such systems respond to these shifts. This was proposed in earlier work in SafeML. This work focuses on the use of SafeML for time series data, and on reliability and robustness estimation of ML-forecasting methods using statistical distance measures. To this end, distance measures based on the Empirical Cumulative Distribution Function (ECDF) proposed in SafeML are explored to measure Statistical-Distance Dissimilarity (SDD) across time series. We then propose SDD-based Reliability Estimate (StaDRe) and SDD-based Robustness (StaDRo) measures. With the help of a clustering technique, the similarity between the statistical properties of data seen during training and the forecasts is identified. The proposed method is capable of providing a link between dataset SDD and Key Performance Indicators (KPIs) of the ML models.