A Framework for End-to-End Deep Learning-Based Anomaly Detection in Transportation Networks

TL;DR

This paper introduces EVT-LSTM, an end-to-end deep learning model that combines LSTM with Extreme Value Theory for improved anomaly detection in transportation network data, outperforming existing methods.

Contribution

The paper presents a novel EVT-LSTM model that integrates EVT principles into LSTM for enhanced anomaly detection in temporal transportation data.

Findings

EVT-LSTM outperforms baseline models in diverse real-world datasets.

The model effectively captures rare extreme events in transportation networks.

Experimental results demonstrate superior detection accuracy over traditional methods.

Abstract

We develop an end-to-end deep learning-based anomaly detection model for temporal data in transportation networks. The proposed EVT-LSTM model is derived from the popular LSTM (Long Short-Term Memory) network and adopts an objective function that is based on fundamental results from EVT (Extreme Value Theory). We compare the EVT-LSTM model with some established statistical, machine learning, and hybrid deep learning baselines. Experiments on seven diverse real-world data sets demonstrate the superior anomaly detection performance of our proposed model over the other models considered in the comparison study.