Learn to Predict Vertical Track Irregularity with Extremely Imbalanced Data

TL;DR

This paper presents a framework for predicting vertical track irregularities using large-scale real-world data, employing advanced machine learning techniques and a novel method to handle data imbalance for improved detection of rare deformation events.

Contribution

It introduces a new approach combining adaptive data sampling and penalized loss to effectively address data imbalance in multivariate time series prediction for railway track deformation.

Findings

Enhanced prediction accuracy for rare irregularities.

Effective handling of imbalanced data improves model sensitivity.

Demonstrated success on large-scale real-world railway data.

Abstract

Railway systems require regular manual maintenance, a large part of which is dedicated to inspecting track deformation. Such deformation might severely impact trains' runtime security, whereas such inspections remain costly for both finance and human resources. Therefore, a more precise and efficient approach to detect railway track deformation is in urgent need. In this paper, we showcase an application framework for predicting vertical track irregularity, based on a real-world, large-scale dataset produced by several operating railways in China. We have conducted extensive experiments on various machine learning & ensemble learning algorithms in an effort to maximize the model's capability in capturing any irregularity. We also proposed a novel approach for handling imbalanced data in multivariate time series prediction tasks with adaptive data sampling and penalized loss. Such an approach has proven to reduce models' sensitivity to the imbalanced target domain, thus improving its performance in predicting rare extreme values.